Matsoft :: Home
biletul zilei cu meciuri din fotbal omnibet biletul zilei de azi la pariuri sportive
 
Home
|  
Companie
|  
Produse
|  
Services
|  
Projects
|  
Research
|  
Presentation
|  
Comanda
|  
Contact
|  
About
|
 


Various Information
MAT SOFT is a brand by which Mircea Digulescu is known in Romania and other countries. The name MAT Soft Technology, distinct from MatSoft is the name of a company he founded in 2009 and sold in 2014.

Mat Soft Technology was a Private Cloud company functioning in the field of Business Process Automation and Robotization (RPA, BPMS). When he sold in 2014, by the terms of the sale, the codebase for that project was branched, with Mircea Digulescu retaining rights to build on top of that state, including for commercial purposes.

The present website belongs to Mircea Digulescu exclusively. It was previously used to promote his Mat Soft Technology project, and some portions of it have not yet been update to reflect recent changes.



top


Contact Us
Contact us using the Contact page.

Below there is a picture of MAT SOFT's founder, Mircea Digulescu.
Mircea Digulescu (at Softbinator Tracing the Roadmap Conference)

top


Presentation of Mircea Digulescu and his Proposals

Mircea Digulescu is an Independent Entrepreneur, Leader, Computer Science Researcher, Software Engineer, Intelligence Enthusiast and Amateur Writer. He is currently looking for partners to contract his services and investors to finance his projects. He is also interested in bilateral and multilateral scientific cooperation (in a private capacity), as well as offering trainings.



He is originally from Bucharest, Romania and currently operates worldwide.


He has a PhD (ABD) in Applied Computer Science from University of Bucharest (2019), Faculty of Mathematics and Computer Science, where he also taught briefly (2014), and was awarded the Professional Diploma in Management (with Distinction) by the Open University Business School (OUBS) in Milton Keynes (distance learning) (2008). He also completed several other graduate and post-graduate courses, including Russian Language for Foreigners from Lobachevski State University – 1 semester (2019), Competitive Negotiation Strategy - Moscow School of Management SKOLKOVO (2013), International Strategy - OUBS (2011) and was offered admission but did not attend Moscow School of Management SKOLKOVO (in 2012) and Cornell University (in 2004).

Mircea Digulescu is a former laureate of the toughest, most renowned international Informatics contests for University students and high school pupils (CEOI 2003, 2004; ACM SEERC 2004, 2005) and is also a Div1 coder on Codeforces.com (profile: mircea85) – peek score 2022 candidate master.

He self-published a number of papers on topics of Computer Science – mainly Game Theory (like Farsighted Stability in Stable Marriage Problem), Complexity Theory (like Towards Solving NP-Complete and Other Hard Problems Efficiently in Practice) and Cryptography (like Hiding Data in Plain Sight: Towards Provably Unbreakable Encryption with Short Secret Keys and One Way Functions) – some of which have been peer-reviewed and published at conference. He also self-published a number of articles on social awareness and strategy pertaining to the topic of designing and enacting a better society: eg. On the Functioning of Society (2019), On Happiness (2015), etc.

Mircea Digulescu founded an RPA technology startup (Mat Soft Technology) in 2009 which he sold in 2014. He worked on various other projects in the field of software engineering, with roles varying form Senior Software Engineer to de-facto Director of Engineering (Engineering Manager officially) to Technical Product Manager, mostly freelance. Up to Feb 2021, he never earned money from anything other than software engineering in his entire existence.

Mircea Digulescu has some experience working with international actors, which he gained by travelling around on business trips for his former startup and then to try to raise funds.

Mircea Digulescu is a fully independent, non-state actor. All his non-software endeavors were and are self-financed. He maintains himself aware of the problems and situations across most regions of the globe, by individually processing OSINT (especially NEWSINT) and the occasional, opportunistic HUMINT he obtains via his personal contacts with others. When he was young (9-14 yo.), Mircea Digulescu developed the system of functional “secret service” where he co-opted some of his school colleagues at the time (1995-2000) – see photo docs here. He has self-perfected his knowledge and competence in intelligence gathering and operations ever since. He considers himself an Intelligence Enthusiast.

As a Software Engineer he builds fast, reliable systems that scale. He built a Google Maps competitor for Bolt.eu which is currently live in production and serves around 100 Million monthly active users and over 1 Billion transactions. He can build systems of larger scale as well.

As an Entrepreneur, Mircea Digulescu proposes 8 ideas for new startups, including a Marketplace, a Cryptography company and Private Cloud. He had his own RPA (Robotization Process Automation) BPMS startup, MAT Soft Technology during the period 2009-2014. He is currently looking for investors and customers for these 8 new startups.

He speaks Romanian (native), English (bilingual, certified by SAT1+2, TOEFL), Russian (medium, ~B1), German (low), French (very low).

Mircea Digulescu is actively looking for partners to:

1.     Contract his strategy / intelligence consultancy services.

2.     Invest in or become customers of his future tech and consultancy startups.

3.     Help attract investment and partner with him in his future tech startups.

4.     Contract his software engineering and management services or pay him for end-to-end R&D of software and non-software projects.

5.     Finance his activities in Computer Science Research.

6.     Contract training and teaching services from him, particularly in the fields of computer science and software engineering.

7.     Engage in direct cooperation with him.

8.     Popularize his works, especially scientific papers as well as his belletristic works as an amateur writer. This includes translations.

9.     Contract Mircea Digulescu to lead private and public organizations.

Contact Mircea Digulescu directly for any of the above opportunities: mircea.digulescu@gmail.com (preferred); mircea.digulescu@matsoft.ro;

The reminder of this document details the portfolio of projects and activities conferring Mircea Digulescu the credentials to offer a successful cooperation in each of the above categories.

The concrete details of types of products, services and companies being proposed are briefly detailed as well. Contact Mircea Digulescu directly for any additional information and pricing.

Some of the credentials for one of the 8 categories apply to others. It is recommended to read this document in its entirety.


top

1. Intelligence and Strategy Consultancy by Mircea Digulescu

Mircea Digulescu has a passion for some activities in the information and strategy sphere. His self-developed and practical experience might thus sometimes out-perform that of some established entities.

The single most important contribution that Mircea Digulescu can bring is setting up and operating as part of a strategic think tank offering deep, useful, actable insights and creative solutions and approaches to concrete problems and operations facing decision makers and fellow strategists. Additional proposals for specific cooperation include but are not limited to the following areas:

·         Resiliency by Design. Building and offering insights into how to create the Structures, Systems and Culture of Organizations and how to scheme them with personnel so as to make them by design – game theoretically – hard to compromise via cooption and infiltration and also how to enable them to function effectively.

·         Operations. Anticipating strategies of adversaries. Analyzing and evaluating treason risk, uncovering infiltrators and agents foreign to some organization. From individual experience and leveraging Game Theory background this can be particularly effective when countering the western methodology of infiltration and compromise. Competitive negotiation support.

·         Cryptography, Information Security and IT Operations. Mircea Digulescu is quite knowledgeable about Cryptography and has invented a novel type of symmetric key cryptographic system himself (see the cryptography part of section 4 below). Other schemes in pipeline include a novel Symmetric Key Authentication scheme over a public channel (with zero leaked knowledge), FoF identification, digital signature systems and cryptographic chain of command authorization systems.  Information Security and ITOPS are also in scope.

·         C4I IT Systems. Developing software system to control aid and generate insights pertaining to C4I operations and maintenance of critical infrastructure in working order. Developing modules or offering services related to such.

·         Technical Software. This is software that runs on embedded, forward deployed hardware.

Cooperation in this sphere can take the form of on-demand or revolving contracts, public speaking, trainings, compiling reports and recommendations and leading concrete projects to completion.

Portfolio of past activities pertaining to Intelligence, Strategy, Diplomacy and State Relations can be found on the Presentation Page.


top

2. Leadership Services by Mircea Digulescu

Mircea Digulescu is both a leader and a manager. He is particularly talented in inspiring and “putting together” groups of smart people. He can also communicate effectively across several different echelons of smartness and awareness on life and world affairs.

Proposals for Leadership Services include:

·         CEO, Board, VP and the like. Proactively leading organizations small or big, setting up and enacting effective strategic direction in accordance with client’s goals.

·         Executive Advisor. Serving in an Advisory decision support role, on a strategic as well as operative level (for a concrete operation). Improving quality of decisions, discovering creative alternatives to hard problems and effectively communicating and gathering stakeholder support for decisions. Following through and verifying their enactment.

·         Offering trainings and consultancy of a general or specific nature. Sharing best practices and educating others on how to approach and handle different situations, both actual and prospective. This task can be performed with or without access to secret information.

Mircea Digulescu has held a number of leadership and managerial roles in his professional career. While the portfolio below speaks for itself, it is less than ¼ of 1% of the potential capacity for leadership that Mircea Digulescu can manifest. The situation is such because, having been born and living for a long time in Romania, due to the socio-political context, he was never entrusted with a significant role.

Mircea Digulescu’s leadership portfolio and credentials include:

·         Founder, CEO and CTO of Mat Soft Technology (2009-2014). Raised 165k euros (from 500k euro initial promise) at a $2.5M valuation from distant relative of his, a Romanian business man and, together with his own funds, created the company in 2009. The startup aimed to offer a private cloud infrastructure, on top of which a business process automation engine (BPMS) can be run and operated. From a commercial perspective this was a Robotization Process Automation (RPA) company in present day terminology. See details about the product and the phase to which it was bought here: http://www.matsoft.ro/index.php?sectionId=108. In 2012, the investor terminated financing being generally obscure as to reasons. Leading Mat Soft Technology involved functioning as both CEO and CTO (having invented the platform and algorithms of the underlying technology) and leading a team of 4-9 elite software engineers as well as handling stakeholder relationships and legal matters. After Mircea Digulescu sold his company to investor in 2014 for 1 euro, the idea was picked up and successful developed by Romanian “unicorn” UiPath.

·         matbs6Director of Engineering. Leading the Bucharest Engineering Office of Bolt.eu and growing it from 3 to 23 people senior engineers (2018-2019). Bolt.eu (then Taxify) is the European ride-hailing and on-demand tech services provider, competitor to the US Uber and Lyft. The official title was Engineering Manager, but the effective work was that of Director. Leadership involved some elite technical contributions (solutions) as well (see Section 3 below), but included holding 1-1 meetings, screening candidates via interviews, performance management and evaluation, appraising migration and other risks for each office staff member, having technical meetings and maintaining relations with the corporate parent (including status updates) remotely via the Head of Engineering.

·         Vice President of Technical Product Management with DevFactory (2018 and 2020). DevFactory is a large IT conglomerate which is a terminus points for many small startups which fail to self-sustain. The role involved discovering, fixing and improving the technical architecture of the products of such companies and finding ways to integrate them into existing portfolio. Work was of an elite level (moderate technical depth though). This was a leadership role because it involved working in an advisory role to company CEO and President, as part of a team.

·         Product Manager with SIVECO Romania (2005-2006). This was my second job at my first employer (20-21 yo). Starting as a software engineer, I was quickly noticed by VP of Engineering and Execs of SIVECO Romania and moved into a semi-leadership role: I owned the technical roadmap of the Business Intelligence and Analytics (OLAP) product of the company. They also used me to do presales and to fix stuff before customer acceptance for big budget gov’t contracts, like The National Bank of Romania and the romanian National Health Inssurance House, as well as pitch the product to private companies on the Romanian market at the time, like Accelor-Mittal group (Steel), Romanceram (Ceramics), Velpitar (bakery) and others. While they made quite some money from me, I was not close to any leader there and was not involved in sharing profits.

·         Professional Certificate in Management, The Open University Business School (2005-2006) – distance learning, via Codecs. Learned MBA and Management (undergraduate level). Was quickly remarked and courted by the non-gov’t secret organizations for the first time.

·         Professional Diploma in Management, The Open University Business School (2007-2008) – distance learning, via Codecs. I continued my MBA studies at a postgraduate level and completed the program with Distinction (top <5%). Course involved both individual work, tutor marked assignments as well as brief in-person meetings in Bucharest and Brussels for residential schools.

·         Moscow School of Management SKOLKOVO (2013). Was offered admission for fulltime MBA program but did not join since it was terminated the same year and also required 60k euro payment. I did have the opportunity to take an interactive demo course for Competitive Negotiation with renowned Israeli lecturer Moty. Application process took place 2010-2012.

·         See further relevant details in the Summary and Sections 1 and 3 of the Presentation Page.


top

3. Computer Science R&D and Teaching by Mircea Digulescu

Mircea Digulescu is an independent researcher in Computer Science. He never got financing for his activities, but is actively looking. His interests and publications are detailed below. He would like to identify partners to finance research and development up to products in each of the following areas:

·         Cryptography. Mircea Digulescu is keen on Cryptography and has invented a novel cryptographic scheme based on a new technique he calls entropy enhancement. While not a mathematician, he is knowledgeable about the workings and fundamental security assumptions of most cryptographic schemes and protocols today. Also he is insightful about implementing and using cryptographic systems securely, avoiding common pitfalls like timing attacks, using low-entropy seeds, dictionary-based edit-distance attacks, using compromised machines to enter the secret key and more.

o    In Pipeline: Functional Encryption: Turing-complete computation over encrypted data using SKREM-like ciphers (~2025?). The aim is to allow computations of arbitrary algorithms for a registry machine (which is Turing complete) over encrypted data, without knowing the decryption key. This would allow for scenarios where data can be uploaded to the Cloud in encrypted form, the cloud can perform some advanced operation without revealing the algorithm to the user, and finally send back the result to the user also in encrypted form. As a result, the Cloud processor never learns the data it processes, and also the end-user never learns the details of the advanced algorithm used. This idea is in development.

o    In Pipeline: Light Reading on Cryptanalysis: automating breaking of substitution ciphers and recognizing plain texts (~2023). The purpose is to build a system which takes as input a text encrypted with a substitution cipher and efficiently outputs the plain-text. As ciphers become ever more sophisticated, this can be useful in brute force attacks on more sophisticated encryption, in cases where not enough of the original plain text is known. Ideas for the paper have been fully developed.

o    In Pipeline: Secure Coin Flipping and Cryptographic Commit Protocol based on SKREM-like ciphers (~2021).  Allows agents to commit to choosing a certain number without revealing it upfront. This in turns allows secure coin flipping. Unlike existing schemes whose security relies on prime factorization at its core, this scheme relies on security of SKREM-like ciphers. Ideas for the paper have been fully developed.

o    In Pipeline: Super Simplified SKREM: Practicalizing Unbreakable Symmetric Key Encryption with Short Secret Keys (~2021). Introduces a simplified, practical variant of SKREM ciphers, based on the paper below, which has better algorithmic performance characteristics (number of random bits required, running time and memory complexity). Ideas for the paper have been fully developed.

o    Completed: Application of SKREM-like ciphers (2021 July). In a prior paper we introduced a new symmetric key encryption scheme called Short Key Random Encryption Machine (SKREM), for which we claimed excellent security guarantees. In this paper we present and briefly discuss some of its applications outside conventional data encryption. These are Secure Coin Flipping, Cryptographic Hashing, Zero-Leaked-Knowledge Authentication and Authorization and a Digital Signature scheme which can be employed on a block-chain. We also briefly recap SKREM-like ciphers and the assumptions on which their security are based. The above applications are novel because they do not involve public key cryptography. Furthermore, the security of SKREM-like ciphers is not based on hardness of some algebraic operations, thus not opening them up to specific quantum computing attacks. See here and here. See peer-reviewed published version here and here.

o    Completed: Hiding Data in Plain Sight: Towards Provably Unbreakable Encryption with Short Secret Keys and One Way Functions (2019 December). Introduces a novel symmetric key encryption scheme based on a technique of entropy enhancement based on the theory of algorithmically random sequences, by requiring up-front a large master table of random bits. Unlike existing schemes, it does not rely on the hardness of algebraic operations (such as the algebraic field inverse) but instead on that of indirection based on a random sequence. This scheme is called SKREM and is claimed to be provably unbreakable. See here or here.

o   Existing Topics Known: AES (Rijandel), RSA, El-Gamal, Kerberos as well as basic Quantum Cryptography, Kolmogorov Extraction, Algorithmically Random Sequences, SHA, MD5, Merkel Trees, elementary Functional Encryption, interactive proofs.

·         Complexity Theory and Computability.

o    In Pipeline: Building Hard Cases for Hard Problems (~2024). Discover (perhaps automatically) cases for hard problems which cannot be easily solved using the techniques of the 2022-2024 papers or any other approaches.

o    In Pipeline: How Intractable is Intractable: Solving Kolmogorov Complexity on the finite case (~2024). Discover (perhaps automatically) an algorithm which proves Kolmogorov complexity of short strings (up to 128 bits long? Or maybe 1024?) – or at least conveniently bounds it. Based on the techniques of the 2019-2023 papers.

o    In Pipeline: Automatic SAT Solver (~2023). Discover (perhaps automatically) and a better 3-CNF-SAT solver than those in existence, based on the ideas presented in the 2019 paper.

o    In Pipeline: Automatic Algorithm Finder (~2023). Refine and Implement the techniques from the 2019 paper to produce an algorithm which finds algorithms to solve hard problems. How fast will it find KMP, constant time 32- and 64- Bit Counting or A*? How about a good algorithm to efficiently factor 50 digit numbers?

o    In Pipeline: Exact 3-CNF-SAT solver when there are many solutions (~2022). Exploiting the fact it is known that there are many solutions to a 3-CNF-SAT expression, finding one should be faster than in the general case.

o    Completed: Towards Solving NP-Complete and Other Hard Problems Efficiently in Practice (2019 November). Introduced the field of Finite Algorithmics: reasoning about problem difficulties for bounded inputs. Offered ideas and approaches for automatic and computer-aided discovery of algorithms for hard problems, on the finite case. Presented concrete directions for solving 3-CNF-SAT, Kolmogorov Complexity and Prime Factorization. See here or here.

o    Completed: Nondecidability of Halting Problem – a Personal Proof (2013 September). Constructed a paradoxical computer program proving that there is no general-case solution for the Halting Problem. See here.

o    Existing Topics Known: Finite Automata, Push Down Automata, Theory of Computing, Amortized Analysis, Master Theorem, Akka-Bazzi method, Intractability, Complexity Classes, NP-Completeness, Arithmetical Hierarchy, Quantum Computing, P-Spiking Systems, Bioinformatics (wrt Computability), Random Turing Machines, etc.

·         Game Theory.

o    In pipeline: Collusion with Binary Objectives in the Stable Marriage Problem (~2023). Continues the exploration of the 2021 paper, with a slightly modified objective function: the boys can either get a particular match or better or otherwise they are fully indifferent as to their individual outcome. Ideas for the paper have been almost fully developed.

o    In pipeline: Resiliency by Design: Using Game Theory to structure organizations to withstand infiltration and treason (~2023). Addresses how to structure organizations into groups and loyalty thresholds based on individual outcome and trust in order to withstand poaching and infiltration. Idea is in development.

o    In pipeline: War Games: Withstanding and Winning Protracted Conflicts using Game Theory (~2022). Deals with the concept of balancing production and spending with active military operations in order to withstand and win conflicts spanning a long period of time (such that war material is constantly being produced and churned). Idea is in development.

o    In pipeline: Mosquito Drone Strike: Attack and Defense versus a Swarm of Drones (~2022). Expands upon the paper below to deal with the case where missiles can travel backwards in space. Idea is in development.

o    In pipeline: Hold the Line: Missile versus Interceptor duels and the k-servers problem (~2022). Deals with identifying the best strategy for a massive missile attack and the corresponding best response on the defense side. Idea is in development.

o    In pipeline: Survival and Prosperity: Multi-Currency Investment Strategy with forced spending (~2022). Deals with determining the optimal strategy for reaching an objective over a period of time starting with an initial capital across multiple currencies and choosing the most adequate investments for this purpose. Idea is in development.

o    In pipeline: Threats and Bribes in the Stable Marriage Problem (~2021). Expands upon the 2019 and 2016 papers to the case where boys are willing to take risks of a worse outcome in order to issue ultimatums to one-another before game play so as to obtain a better match in case their threats work. Paper discusses the general structure of this new ultimatum problem and proposes an algorithm for identifying a reasonable strategy under a specific model. Ideas for this paper have been fully developed.

o    In pipeline: Applications of the Stable Marriage Problem (~2021). Introduces a series of problems which can be formulated appropriately in terms of Stable Marriage Problem, including the likes of Vendor versus Contracting Authority, where an oligopoly of vendors colludes to get the best prices and projects from a number of accepting customers. Ideas for this paper have been fully developed.

o    Completed: Weakening Nuclear Posture can result in added security (2019 Feb). Brief note on how certain combinations of doctrines lead to war, while weakening that of just one actor can result in peace. See here (in Romanian).

o    Completed: Farsighted Collusion in Stable Marriage Problem (2019 May). Presented an efficient algorithm for computing the unique farsightedly stable boy-optimal matching in the Stable Marriage Problem. Also criticized earlier work pertaining to the subject and presented a few connected interesting results. The large body of the work is an independent rediscovery of results by Jun Wako (2010), adding upon them further details (a full implementation) of the efficient algorithm and structuring the proofs in a language more attuned for computer scientists (versus mathematicians), requiring fewer theoretical prerequisites. See here, here or here.

o    Completed: Strategic Play in Stable Marriage Problem (2016 Aug).  See here or here. Discussed in a very rough form a host of results pertaining to the Stable Marriage Problem, including applications thereof.

o    Completed: Dealing it multiple times does not change the expected value (2015 May). Discusses how the common practice of dealing the turn and/or river multiple times in holdem poker games does not change the player’s expected value. Implies that random samplings of this type can be performed without substitution (applicable in A/B tests). The proof involved two innovative steps. See proof here and here and code sample here and here.

o    Completed: Project: Success Probability Calculator (2015 April). Presents the implementation of an O(n*m) method for identifying the optimal investment strategy to achieve a target single-currency bankroll starting with some capital, choosing one of m investment options over n time periods. The project involved using of some more advanced data structures such as an efficient LRU cache implementation and a Space vs. Time Tradeoff Structure. See details here and source code here or here. The code sample includes an efficient LRU Cache implementation and other helper structures.

o    Completed: Choosing a Gamble with Best Expected Value is often a bad strategy in Stochastic Games (2014 Aug). Argued by example that this is the case. See details here (in Romanian).

o    Completed: A Game of Ultimatum – Surprising Results of Rational Play (2013 Dec). Inspired by a 2013 lecture by negotiation expert Moty from Moscow School of Management SKOLKOVO, a particular 3-player general-sum game of ultimatum was examined, leading to surprising results. See details here (in Romanian).

o    Existing Topics Known: Two/Many Player Zero-Sum Games, General Sum Games, Kernel of a Game, Nash Equilibrium, Discrete Stochastic Games, Sprague-Grundy Numbers, some intractability results, etc.

·         Algorithms and Data Structures.

o    In pipeline: Adaptive Topology in Distributed Solving of Sorting Problems (~2023). Introduces a O(m*log n) network time + O(n*log^2 n) external time method for distributed sorting, involving a dynamic, adaptive network topology. While the O(m*log n) bound is matched much simply by the MM algorithm using sorting networks, the approach of adaptive topology is judged to be of general interest and the method is presented as an inspiration for other researchers to examine this kind of approach. Ideas for this paper have been fully developed.

o    In pipeline: Sub-linear Time Routing in Urban Networks (~2023). Introduces an efficient method, based on smart precomputation to support queries of the form “shortest path from A to B” in graphs representing Urban Networks, which should run much faster than Dial’s algorithm and A*.

o    In pipeline: Memory Efficient Van Emde Boas Data Structure (~2023). Combines a vEM Tree with a Universal Perfect Hashing scheme to obtain an expected time O(log log U) INSERT / DELETE / FIND / UPDATE / MIN / MAX / PREDECESOR / SUCCESOR data structure with only O(n) memory for n queries of keys in the universe 0..U. This improves on vEM Trees in terms of memory consumption from O(U) to O(n).

o    Completed: Shortest Path in a Weighted Graph with Small Edge Costs (2003, GInfo). An independent rediscovery of Dial’s algorithm O(V+E*C) and the brief presentation of a method to further reduce complexity to O(V+E*log log C) using a vEM Tree. Neither Mircea Digulescu, nor the reviewer from GInfo, were aware of Dial’s algorithm at the time (which is so obscure in literature that finding it on Google Scholar is still unsuccessful as of Feb 2021).

o    Completed: Kirkpatrick’s Algorithm (2003, Ginfo, with Andrei Matei coauthor). Presented the famous Kirkpatrick Algorithm and Data Structure for efficient point location in the plan using successive triangulations.

o    Other independent rediscoveries: Kosaraju’s Algorithm, Fisher Yates Shuffling Algorithm.

o    Existing Topics Known: Mircea Digulescu is a recognized as an expert on Algorithms and Data Structures. He is familiar with the following: AVLs, KD-trees, Interval and Segment Trees, B-trees, AB-trees, Orthogonal Range Queries with Fractional Cascading, RMQ, LCA, Binary Heaps, Binomial Heaps, Fibonacci Heaps, Universal Perfect Hashing, Disjoint Sets, KMP, Rabin Karp, Dijkstra, A*, Bellman-Ford-Moore, Floyd-Warshall, Dial, Egg Dropping, Network Flow – Floyd-Fulkerson and Lift to Front, Paxos and many, many others. Also, besides knowing and understanding these algorithms (some of which he independently discovered while reading Cormen CLRS when he was 15-17 yo), he is well apt to apply them, with a vast industry experience and over 320 problems solved on Codeforces.com, where he is a Div1 coder: mircea85. Additionally he solved many non-research problems from the Cormen CLRS and Knuth TAOCP computer science books.

Besides being interested in obtaining financing to continue research as per his interests above, Mircea Digulescu is also interested in partnering with someone for coauthorship publication (in recognized peer-reviewed journals), Translation of results in another language and Dissemination of said to appropriate recipients, with proper attribution.

Mircea Digulescu is also apt to teach or train on topics including advanced Computer Science, both based on his own research interests and results, as well as at the Master’s (for example covering the material of MIT Advanced Data Structures) and at the Bachelors’ levels (for example covering MIT Introduction to Algorithms or the entirety of Cormen CLRS book, including exercises). Additionally he can teach or train on Software Engineering, including writing clean and maintainable code, fundamental DevOps, code reviews, architectural and design patterns and building or using Cloud services, like AWS SQS or Redis.

Mircea Digulescu’s computer science and education credentials include:

·         Div1 coder on www.codeforces.com  roughly in Top 0.3% best of the best coders of all time world wide - http://codeforces.com/profile/mircea85 - highest score 2022, candidate master, with over 320 problems solved (NB: I continue training here, so current rating can vary).

·         Codility Golden Award for the Calcium 2015 Challenge2015 – see here.

·         10th place at the ACM ICPC Southeastern Regional Contest - 2005

·         4th place at the ACM ICPC Southeastern Regional Contest – 2004

·         1 Bronze Medal at European International Olympiads in Computer Science (CEOI) - 2004

·         1 H. Mention at European International Olympiads in Computer Science (CEOI) - 2003

·         3rd prize at the Romanian National Computer Science Olympiad – 2004

·         3rd prize at the Romanian National Computer Science Olympiad - 2003

·         Multiple prizes at other lower-level Computer Science and Mathematics Olympiads and contests

Teaching Experience and Education:

·         University of Bucharest - Faculty of Mathematics and Computer Science – Doctoral Studies in Applied Computer Science – 2014 - 2019 PhD (ABD) completed. Main contributions related to Stable Marriage Problem and to Distributed Sorting. I encountered heavy resistance at getting my work published in indexed peer-reviewed journals – for matters of form - not lack of correctness or relevance, which lead to delays. Finally, in summer of 2019, after completing all work except dissertation, I withdrew from the PhD program in order to get my diplomas. I thus graduated ABD – all but dissertation (for which indexed publications were required).

·         Trainer at Academia Credis (IT Training) –Java Basic (2017): I taught the Java Basic course to a class of about 15 adult students as a side job, more out of passion and my desire to learn Java as well. This was an on-site, in person project, during evenings. I earned 1000 EUR for the month-long course.

·         Assistant Professor at University of Bucharest - Faculty of Mathematics and Computer Science ( Romania’s top University - http://fmi.unibuc.ro/) (2014): I taught Object Oriented Programming (OOP, C++) and Formal Languages and Automata (LFA) courses, holding lectures, providing hands-on assistance, creating and handing-out workouts, providing written and oral feedback and generally developing specified aptitudes. I had a great time and all of my students got excellent industry positions later on. I had a great time teaching. I earned about 350 Euros / month while working here.

·         Train of Trainers Course (2014) at Frends. Completed this program at the invitation of a friend at the time. Frends is a volunteering non-profit, associated with University student groups and student life.

·         National Students’ Congress (2013) – Palace of Parliament (Bucharest) – Speaker. I blabbered a bit about some irrelevant educational topics. Was invited there by surprise.

·         Softbinator Tracing the Roadmap Conference (2012) – Romanian National Library (Bucharest) – Speaker. I spoke on topics of Entrepreneurship and explained the value in having a vast number of autonomous economic actors, versus dealing with an oligopoly.

·         University of Bucharest - Faculty of Mathematics and Computer Science – Master’s Degree Software Engineering (2008-2010). Dissertation thesis “Resource Sharing in a multi-process, multi-transaction environment”, Merit scholarship during 1st year.

·         University of Bucharest - Faculty of Mathematics and Computer Science Bachelor’s Degree Computer Science (2004 – 2008). License thesis “CryptoSafe” – a WinRar desktop application which encrypts data with AES instead of compressing it. Merit scholarship during 1st year.

·          National Computer Science High-school “Tudor Vianu” Baccalaureate, High-school Degree – (2000 – 2004). During high school I joined fashion of taking SAT tests (1400 SAT1, 1800 SAT2 Math IIC, 1730 SAT2 Physics) and applying to US University, and I was offered admission at Cornell University but never joined - declined eventually due to financials mainly.

·         See further relevant details in Section 2 of the Presentation Page.

See Mircea Digulescu’s publications and profile on Academia.edu here and on arXiv.

Mircea Digulescu also has the competence to teach or train on topics of Leadership, Strategy, Intelligence, Business Administration, Organizational Architecture, as well as others.


top

4. Software Engineering by Mircea Digulescu

Mircea Digulescu is a software engineer, engineering leader and manager. He has ample experience with a number of customers across different verticals (Finance, Government, FinTech, OnDemand, Media, Healthcare and, of course, Tech). He is easy to interact with from the customer side, including beneficiaries, managers, executives and owners, autonomous and effective at understanding and clarifying requirements, even in the presence of ambiguity and can lead teams to speed up the software development process. Most of his customers approach Mircea Digulescu for the following:

·         Building systems that scale. Performance and resiliency are key aspects of systems developed. Choosing the proper system architecture, implementing (and potentially discovering) the proper algorithms to apply are key to the success of such a project. Mircea Digulescu can build systems that serve more than 1 Billion monthly active users with over 1 Trillion transactions per month. He can also build near-real time and – with proper hardware – real-time systems.

·         Building substitutes for Cloud technology or Expanding existing Cloud technology. Mircea Digulescu has the technical know and is savvy to rewrite and expand / improve products which substitute existing Cloud services, such as AWS SQS, SNS, S3, Redis, SQLLite and many others. He can rebuild such services from absolute scratch, or use existing open source repositories (eg. Kafka, Redis) to speed up delivery. He can also build non-standard components, suited for a particular use. This is especially handy for building Private Clouds and for solving concrete engineering problems for organization which do not want to or cannot host their applications on public clouds or use open source technology (for example for security reasons).

·         Solving difficult and open-ended (undelimited) engineering problems. Besides run-time performance, many companies require that problems for which it is inherently difficult to arrive at a solution be solved. For example, for Bolt.eu (European Uber/Yandex Taxi), Mircea Digulescu developed the technical architecture to provide them with their own version of Google Maps, including deducing traffic patterns and congestion from existing GPS data from their drivers on the roads. This allowed the company to save many millions in annual costs and gain strategic independence from Google. He did this based on the paragraph-long requirement: “We need to have an independent alternative to Google for routing.”

·         Performance and Accuracy Optimization. For companies that already have existing products or code-bases, Mircea Digulescu is usually called upon to make them actually work – either making them run acceptably fast (or blazing fast as a competitive advantage sometimes) or improving the accuracy of the data they produce or the way they function.

·         Big Data Analysis and Process Automation. When data volumes get large, most “standard” software engineers become overwhelmed and try to improvise some shaky system using open source components, coupled together in way that either doesn’t work, or doesn’t work well enough. Mircea Digulescu offers and implements solutions that live up to the required performance, reliability, automation, accuracy and ease of use requirements, while leveraging existing investments as much as possible.

·         Building Reliable and Resilient Systems (Enterprise or otherwise). When a “standard” system needs to be built with confidence that it will function as desired, be easily scalable in the future and adequate, Mircea Digulescu can architect and implement it, confidently within a fraction of the time it would take others to do this at an acceptable level of quality, should such hypothetical others ever succeed.

·         Fixing the hard stuff. Sometimes even systems with a long history of running in productions have undesired traits (bugs) which have baffled engineers for a long time and which have since become critical to fix. Mircea Digulescu can save the day in such a time.

·         File:SDLC - Software Development Life Cycle.jpgFull Technology Lifecycle – DevOps from inception to decommissioning. Software systems, like most hi-tech systems (eg. ICMBs), have a lifecycle which goes far beyond the first production launch. It includes doing patches, upgrades, migrations, integrations and so on, all while maintaining the system operational and ensuring the integrity of the data. Organizations can get all these problems off their minds, by delegating ownership of the entire lifecycle to Mircea Digulescu. This includes human interactions with technical teams and team-leads/managers as well as IT systems.

·         Managing Engineering Teams. Mircea Digulescu can create and leverage an effective tech culture (similar to big companies like Google, etc.) where people are excited about and productive in their work. Also, he can adjust and perfect such a culture to apply to the demands of a startup – including cutting corners for faster TTM, handling morale, keeping sponsor execs up-to-date and in the loop, and many others.

Mircea Digulescu considers himself a language agnostic programmer, with a dual Computer Science and Engineering background, apt to work on any tech-stack suitable or required for a project or product. He already has professional (paid) experience with the following tech-related stuffs: C#, C/C++, .NET, TypeScript, JavaScript, Java, SQL, Microservices, BlockChain, AWS SNS, SQS, ElastiCache, RedShift, CloudFormation, MSSQL Server, Oracle, MySQL, Redis, RabbitMQ, Spark, GIT, Git Workflow, BitBucket, Cryptogprahy, TCP/IP, HTTP, REST, Cognos, Oracle OLAP, Informatica (BPMS tool), Microsoft Teams, JIRA, Jenkins, ASP.net, WinForms, WCF, MVC, MVVM, Kibana, Graphana, Docker, Akka.net, ReSharper and many others. His best-preferred programming language is C# and his second-best preferred is TypeScript. Mircea Digulescu builds maintainable, extensible systems, emphasizing code quality, component reuse and clarity.

Mircea Digulescu is also considering expanding into Machine Learning, AI as well as embedded systems.

Mircea Digulescu’s portfolio of past projects includes, but is not limited to the following:

·         Bolt.eu (then Taxify.eu) https://bolt.eu (2018-2019). Bolt.eu (then Taxify) is the European ride-hailing and on-demand tech services provider, competitor to the US Uber and the Russian Yandex Taxi. The personal technical contribution for Bolt.eu included determining the architecture of system to function as an internal competitor to Google Maps for routing drivers to riders and computing Estimated Time of Arrival (ETA), based on OpenStreetMap and the GPS location points collected from the drivers, and performance-optimizing the Spark jobs part of the dataflow. It also included writing the Redis scripts for fast computation of a location suggestion service (guessing where the riders wanted to go) based on several signal sources and heuristics, obtaining O(1) performance and 95% accuracy. The third important task was writing from scratch a driver-rider matching module, emphasizing performing as many mutual acceptable matches as possible, using the Min-Cost-Max-Flow algorithm based on Ford-Fulkerson method. Additional projects included corrected technical decisions for a compliance reporting project for the Australian gov’t and another data intelligence one for marketing. These are all production systems, running currently with over 100 million active users, and over 10 Billion monthly transactions: when you open the Bolt.eu app to order a ride, the automatic destinations suggestions, the ETAs you see for the pickup and drop-off, as well as the very driver which is assigned to you, all go through systems technically built by Mircea Digulescu.

·         Collective2 https://trade.collective2.com (2017-2018). Collective2 is a “poor man’s hedge fund”, democratizing investing by allowing small individual investors to discover and automatically duplicate – for a fee –, strategies of top traders on capital markets, especially stocks and derivatives. I developed the entire reporting infrastructure for them, including retrieving, caching and serving historical Stocks, Derivatives and FOREX quotes, determining VaR MTM value of portfolios, integrating live quotes, all culminating with an accurate and responsive P&L service. All this was technically interesting work, involving SQL query design and optimization, Database Connection Pooling, Cache Eviction Strategies, Service Monitoring, as well as financial domain specific knowledge, useful for implementing an Option Settlement Calculator and many other details. The work was particularly challenging since the client was a startup, and thus did not want to take the costly alternatives using high-priced historical and live quotes services that most hedge-funds and banks take. While the title was Senior Software Engineer, I was working directly with the company founders.

·         Hearst https://www.hearst.com (2019-2020). Hearst is a western international media conglomerate, the creator, among others of the blockbuster western film The Avengers. My project was for their IT Services department. It involved fixing, modernizing and automating some workflows, mainly pertaining to flow of financial data concerning reimbursements from PeopleSoft into Concur and then Cognos. The project entailed creating a new SQL data model and C# tasks, all while working with stakeholders, users and third parties to correct and remake the flow properly. Automation tools like Informatica were also employed. My title was Technical Architect.

·         Cegeka https://www.cegeka.com (2014-2015). I did performance optimizations and stability fixes for their Cegeka Care healthcare / nursery homes product mini-ERP system, which the existing team was unable to tackle after months of trying. In doing this I reduced the web system’s loading time from >1 minute to 2-3 seconds, and the per-page loading time of heavy pages from 30-45 seconds to about 4 seconds. An interesting part of the technical solution involved implementing an Euler walk of a housing units tree to allow O(1) is-ancestor queries, directly from SQL. Project was in C#/MSSQL/ASP.net. My title was Senior Software Engineer. Project was on site.

·         Societe Generale https://www.societegenerale.com (2016). I worked on a desktop application which is used by traders to trade on markets, especially stocks and derivatives. I implemented a filtering module for financial instruments and provided the technical solution for an efficient Options’ financial curves evaluation module which integrated with their VaR portfolio evaluation module. I also did other fixes and enhancements for this Front Desk trading tool, which allowed traders to perform both manual and automatic tasks, such as hedging, automatic trading, etc. Even now, billions of euros in trade goes through the modules I have built.

·         Finastra (then Misys) https://www.finastra.com (2017). Finastra (then Misys) is a global firm developing software for the Banking System, having more than 90% of the top 50 banks as customers. My responsibilities involved performance optimizations and stability fixes on their core Summit product. Received excellent feedback from both customers and management. I learned a lot about core banking operations, including financial products for Money and Capital Markets.

·         BNR (working for SIVECO) https://www.bnr.ro (2005). While working for SIVECO Romania I did an Analytics and Business Intelligence project from scratch for the Romanian National Bank. The project involved collecting, processing and serving data for two main areas: financial supervision and banking stability. I was involved in operating the 2005 stress-test of the banking sector which identified banks vulenarble to financial shocks such as the upcoming 2008 financial crisis. I also got acquainted with modern money mechanics and the then-in-effect Basel II standard.

·         CNAS (working for Siveco) http://www.cnas.ro (2006). While working for SIVECO Romania I did an Analytics and Business Intelligence project from scratch for Romanian National Health Insurrance House. It involved mainly analytics and projects over consumption of medicine, both for planning and fraud detection purposes. I learned that the Romanian population had overall a rather large consumption of medicine. Project was on site, during the period of time when I was Product Manager at SIVECO Romania – they generally used me to show credibility to the beneficiary and streamline signing of acceptance papers for the >$10 million multi-annual software project.

·         matbs6Mat Soft Technology http://www.matsoft.ro (2009-2014). While bootstrapping and benefiting from a small investment, I designed the technical architecture and implemented a few modules in an ambitions Private-Cloud integrated system, commercialized as a BPMS tool. I was also technically leading a team varying from 3 to 10 people. From a commercial perspective we were a Robotization Process Automation (RPA) company. See details about the product and the phase to which it was bought here: http://www.matsoft.ro/index.php?sectionId=108. The concept is largely the same as that of later successful Romanian “unicorn” UiPath https://www.uipath.com and US intelligence company Palantir https://www.palantir.com. The technical work was pretty advanced, but the small size of the investment did not allow it to be all completed in that go. I worked as CTO, Chief Architect and Software Expert essentially; all while fulfilling CEO duties as well. Project was in C#.

·         DevFactory https://devfactory.com (2018 and 2020). DevFactory is a large IT conglomerate which is a terminus points for many small startups which fail to self-sustain. The role involved discovering, fixing and improving the technical architecture of the products of such companies and finding ways to integrate them into existing portfolio. Work was of an elite level (moderate technical depth though), focusing on the AWS cloud services, like AWS SNS, SQS, SendGrid, ElastiCache, Cognito, Secrets, CloudFormation, etc. One project involved architecting a responsive Contact Tracing App system scaled globally (billions of users), which can be used during pandemics or wrt to people handling secret information.

·         Mircea Digulescu also received Senior Software offers from Google (2011 and 2021) and Amazon (2019) – which is a true attestation of his skills. He declined however, since they didn’t pay enough and imposed geographical restrictions.

·         See also: Menus Demo Project. This is a project done a while ago as a Demo to demonstrate my full stack skills. See source code here and here. Mircea Digulescu’s GitHub ID is mircead85.

·         See further relevant details in Sections 2 and 3 of the Presentation Page, including University credentials.

The single most important thing all his prior customers and beneficiaries would say about Mircea Digulescu is that he delivered where others didn’t; that they can entrust him with a problem and have the confidence he will get it done (i.e. “fire and forget” approach).


top

5. Societal Architecture Research by Mircea Digulescu

Mircea Digulescu wrote and writes on topics of societal architecture, both prospective and current. His views can be approximately described as leftwing republican libertarianism. He is against assassinations. He supports strategic armaments, especially nuclear, but is against biological weapons.

Mircea Digulescu considers the state an alliance of people, much more than a monopoly on physical force. His views core values, both personal and in terms of design of a better society are Truth, Liberty, Justice, Tolerance and Love.

The concrete materialization of Mircea Digulescu’s ideology, particularly in the context of actual practice, given the current state of world affairs is still the subject of ongoing research by himself.

Mircea Digulescu’s social writings additionally aim to educate the youth into properly understanding their socio-economic reality and empowering them with the competence to detect and avoid manipulation and informative intoxications. It aims to give them the core societal skills “that no one teaches them in school”, to strengthen their cognitive and emotional autonomy and nurture their lucidity.

See the list of articles are published by Mircea Digulescu or in pipeline, pertaining to societal topics in Section 5 of the Presentation Page

Mircea Digulescu is doing these writings as a hobby.


top

             6. Direct Cooperation

Mircea Digulescu can function as part of a strategic think tank, or he can directly lead an organization tasked with solving problems technically, at a local, regional, district or central level.

He can might make strategic and IT contributions, similar to the US start-up which developed overwhelming tactic for air-to-air dogfight combat between jets using Machine Learning and AI principles, defeating all human opponents, including some with decades of combat experience.


top

             7. Just for fun writings

Mircea is to be a SciFi writer, leveraging inspiration from his own life in his works. His best target audiences are mainly educated people or people capable of completing higher education but not only. His style is original, with slight influences from James Clawell, Alexandre Dumas and Donald Knuth.

He already wrote a belletristic mathematical work called “Explanation of Love” (2013) – in Romanian, which was rather well received by its target audience. See here, here and here.

His second and main work will be a SciFi love novel, inspired by some of his real-life adventures and thoughts, called “The Little Big Game of Life”: a 3-part story about a He and a She who love each other and are torn apart by the world in which they live, gradually discovering themselves and their reality and eventually ending up saving the Universe a total number of two times. The fictional universe of the book will gradually change from present day mundane to something like that of Star Trek The Next Generation and beyond.

He does this as a hobby.


top

8. Start-up Projects by Mircea Digulescu

Based on his comprehensive experience and technical competence Mircea Digulescu is confident to propose setting up startups on the several ideas. He is currently looking for customers and financing up to self-sustainability from sales for these startups. Additionally he is looking for partners in complementary areas of activity for mutual promotion and creation of synergies. Ideally, he will found all these startups, as part of an umbrella group of his, since they all have merit. He is in it for the long run. Ideally, he would be leading all these companies at the strategic level and as CTO, doing of them in parallel, should raised financing allowed for proper additional staffing.

See the full list of projects on the 
Startup Projects page.

 

Mircea Digulescu is currently actively seeking Seed and Series A financing for each of the above startups. He is also looking for a strategic partner who can support financing for setting up the umbrella group to do all these. Additionally he is actively looking for potential customers which are interested to partner with or acquire software and services from any these future companies.

See Sections 2, 3 and 4 above for further details on what recommends Mircea Digulescu as an entrepreneur and leader for making these startups into a series of successes.


top

The Emotional Part of the Story

Mircea Digulescu knows he is not perfect. He continuously strives to improve one self and show modesty besides skill in all occasions. He understands he can be controversial at times and, while not forgetting about himself, seeks accords which are desired by all parties involved. Nevertheless, he considers the potential for cooperation and achievement by far outweighs any potential challenges by several orders of magnitude.

Mircea Digulescu is also interested to work together with his clients on any kind of related initiatives they may have in any sphere. Please schedule a call or a meeting now by contacting Mircea Digulescu at:

·         Email: mircea.digulescu@gmail.com (preferred); mircea.digulescu@matsoft.ro.

NB: MAT SOFT is a brand by which Mircea Digulescu is known in Romania and other countries. The name MAT Soft Technology, distinct from MatSoft is the name of a company he founded in 2009 and sold in 2014.

Perhaps the main reason why we are chosen over our competitors lies in our motto: We succeed where others fail; we deliver where others deliver excuses. While delivering is the single most important aspect, we also emphasize good communications and friendliness in our relationship with beneficiaries.” – Mircea Digulescu (Managing Partner)

 


top