Country of Origin: Moldova

Education:
M.Sc. Computer Science with distinction (April 2024),
Technical University of Munich,
B.Sc. Computer Science (October 2020),
Technical University of Munich

Languages: Romanian (native), Russian (native), English (proficient), German (proficient)

WORK EXPERIENCE
Deciduous Works, Founder (Jun 2026 - present), Indie Game Development Startup (Multiplatform)

ocell GmbH, (Fullstack) Software Developer (Jun 24 - Aug 26), Forestry Software Startup
• Led and contributed to major features within the web and mobile forest management software Dynamic Forest
(Flutter, Dart, TS) and actively collaborated with the product team
• Architected and led major refactoring efforts to improve performance, code quality and align with DDD principles.
• Mentored a Junior Engineer and regularly shared best practices with the team.
• Led several efforts to improve internal processes and tooling.

Find the Deinotherium, Lead Unity Developer (Jan 24 - Sep 25),
AR Game for kids and teens visiting the Ethnological Museum of Moldova developed with Racketa Production
• Implemented spatial tracking through the museum rooms (Vuforia, Image Targets)
• Implemented the game logic and user interface, based on animations and designs provided by colleagues (Unity)
• Published the app on the App Store and Google Play

Uncountable Inc., Fullstack Software Engineer (Sept 21 – Sept 23),
Data Management and AI Platform for R & D companies
• Created various platform features for displaying and processing data (React, Typescript, Redux, Python Flask)
• Refurbished some of the most-used platform pages (JS to TS, new API calls, code refactoring)
• Contributed to developer tools and IDE plugins to enforce clean code practices
fortiss GmbH, Munich: Data Scientist, (Jun 18 – Jul 20)
• Contributed to creation of a Technology Knowledge graph with 500k items
• Trained and implemented NLP (Word2Vec, cooccurrence-based) models, based on scraped public Big Data
(Python – gensim, nltk, spacy, sklearn, numpy, MongoDB)
• Generated over 1 mil. relationships based on those (Neo4j, Cypher)

Technical University of Munich, Teaching Assistant (2019, 2021, 2024)
• Courses: Efficient Algorithms, Parallel Programming, Numerical Programming, Algorithms & Data Structures,
Discrete Structures
• Prepared theoretical and programming exercises for students
• Held weekly exercise sessions for groups of 10-50 students and checked exams

ACADEMIC WORK
Strategyproof Committee Voting Rules for Restricted Domains of Preferences (Oct 23 - Apr 24)
Master’s Thesis in Computational Social Choice (Grade: 1.0)
• Studied strategyproof multiwinner voting rules in the party-approval setting restricted to candidate-interval
preferences
• Proved that a party-approval voting rule in the CI domain, satisfying strategyproofness and 3 additional axioms is
a composition of phantom median rules, thus extending the result by Moulin (1980)
• Characterized a core-stable subclass of the aforementioned phantom median rules

Detection and Classification of Historic Paper Watermarks with the German National Library Apr 23 - Aug 23
Winner of the poster track at the DhD 2024 conference
• Created a dataset of 5000 data points of paper watermarks and outlines
• Trained CycleGANs for watermark to sketch translation and paired cGANs for preprocessing
• Achieved a 68% accuracy of a correct recommendation within a top 50 scope from 5000+ images