Stefan Dumitrescu
ML EngineerAI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
Head of AI Modeling | Principal ML Engineer
TL;DR: ML / NLP & Systems design My core focus on applied NLP: how to use models to solve real problems, and how to take them from concept to production. I've worked alone, I've led teams; I've been in the code trenches debugging and also on high-level teams working on system architecture. I've played with everything from the first LSTM networks to LLMs, VLMs and Computer Vision models. I enjoy the occasional Knowledge Graph design as much as anybody. I'm contributing to open source and the #AIRomania community, where I've led the development and release of the first Romanian BERT, T5, and GPT-Neo models, plus several foundational NLP datasets. I've also organized an ML hackathon. PhD in Computer Science, back in a time where you had to code everything from scratch.
Université Paul Sabatier Toulouse III
PhD, Information Extraction, Semantic Technologies, Natural Language Programming
January 1, 2008 – January 1, 2011
University POLITEHNICA of Bucharest
MSc, Computer Science
January 1, 2003 – January 1, 2008
Glass, Lewis & Co.
Head of AI Modeling
January 1, 2025 – Present
Valencian Community, Spain · Remote
Adobe
Senior Machine Learning Engineer
March 1, 2023 – June 1, 2024
Adobe
Senior Machine Learning Engineer
January 1, 2020 – February 1, 2023
Sustainalytics
Machine Learning Solutions Architect
October 1, 2017 – January 1, 2020
Bucharest, Romania
Research Institute for Artificial Intelligence, Romanian Academy (RACAI)
Scientific Researcher (CS3)
October 1, 2011 – November 1, 2019
Bucharest, Romania
INEOSOFT
Founder, Software Lead
May 1, 2005 – October 1, 2011
Bucharest, Romania
DRuKoLA - Romania/Germany bilateral project - Team Member
January 1, 2015 – January 1, 2017
The DRuKoLA project that is centered around the German Reference Corpus DeReKo and the Reference Corpus of Contemporary Romanian Language CoRoLa has started in January 2016 and is a cooperation between the University of Bucharest, the Institute for the German Language, and research institutes of the Romanian Academy in Bucharest and Iaşi. DRuKoLa is a transdisciplinary project involving corpus linguistics, computational linguistics, applied linguistics and cross-linguistic studies, applied computer science, corpus architecture and finally also research infrastructure development. My main task is to handle the conversion of text/metadata/annotations to the i5 (Korap) format used by the query platform. Several text processing algorithms are involved, including fuzzy word alignment using Dynamic Time Warp, sub-corpora creation based on automatically balanced metadata, etc.
ANVSIB - PN II national project - Director of the RACAI team
October 1, 2014 – September 1, 2017
The project "Assistive Natural-language, Voice-controlled System for Intelligent Buildings" aims at creating an interface between humans and the smart devices around, using speech as the control interface, with synthesized voice as audio feedback. The project integrates Automatic Speech Recognition, Natural Language Understanding, Text To Speech Synthesis technologies and an industry-standard KNX device communication protocol embedding. RACAI's task is to build the Speech Synthesis and the Natural Language Understanding Modules, as well as varied software components interconnections.
CoRoLa – Priority Project of the Romanian Academy - Team Member
January 1, 2014 – January 1, 2017
RACAI and UAIC (University "Alexandru Ioan Cuza" from Iasi, Romania) are collaborating on the creation of a reference electronic corpus of contemporary Romanian language, i.e., a collection of (written and spoken) texts of great dimension (hundreds of millions of word forms), annotated with metadata (style, author, etc.) and with linguistic data (part of speech, grammatical categories, syntactic dependencies, etc.). My responsibilities in this project are to gather, sort, correct and annotate (both with metadata and text processing) the Romanian corpus. I also created a web-crawler that automatically gathers text resources from a set of predefined sites, as well as an automated pipeline that spans the entire processing chain from raw text to final indexable annotated XMLs.
MetaNet4U - FP7 international project - Team Member
February 1, 2010 – January 1, 2013
The central objective of the METANET4U project is to contribute to the establishment of a pan-European digital platform that makes available language resources and services, encompassing both datasets and software tools, for speech and language processing, and supports a new generation of exchange facilities for them. I have worked in RACAI's team as main lead to collect and verify Romanian language resources, as well as being the main quality evaluator for all language resources and tools in the project.
Machine Learning
Coursera
June 24, 2026 – Present
Probabilistic Graphical Models
Coursera
June 24, 2026 – Present
Deep Learning Nanodegree
Udacity
June 24, 2026 – Present
Statistics One
Coursera
June 24, 2026 – Present
Web Intelligence and Big Data
Coursera
June 24, 2026 – Present
Cultural Fit Analysis
The candidate demonstrates a strong cultural fit for a research-oriented and innovative environment, evidenced by extensive involvement in national and international research projects (CoRoLa, DRuKoLA, ANVSIB, MetaNet4U) and a scientific researcher role at the Romanian Academy. The diverse project portfolio, ranging from corpus linguistics to smart building interfaces, shows a broad interest and ability to adapt to different problem domains within AI/ML. The transition from academic research to industry roles at Adobe and Sustainalytics, and then to a leadership position at Glass, Lewis & Co., indicates a blend of theoretical depth and practical application, which is valuable for an ML Engineer role that often bridges research and product development.
Soft Skills & Operational Fit
The candidate's experience as 'Head of AI Modeling' and 'Director of the RACAI team' suggests strong leadership, project management, and team guidance abilities. Descriptions of designing solutions for internal needs and acting as a product owner indicate a proactive, problem-solving mindset and operational fit for driving ML initiatives. The entrepreneurial background also points to adaptability and resilience.