Dan Smart
Principal AI Specialist at Alexandra Instituttet
Copenhagen, Capital Region of Denmark
Summary
🤩
Rockstar🎓
Top SchoolDan Smart is a Principal AI Specialist based in Copenhagen with a PhD in Mathematics and seven years of cross-sector experience applying ML to NLP, speech and graph problems. He blends academic rigor from postdoctoral research on graph neural networks with hands-on consulting, having led evaluation and model-development work packages for EU projects like TrustLLM and CoRal. His work produced state-of-the-art Danish speech recognition models and helped build the EuroEval benchmarking framework, earning Open Source Award nominations. At Alexandra Instituttet he steers project directions and mentors specialists while shipping production-ready NLP solutions for low-resource languages. He contributes to open-source ML tooling—improving structured text generation compatibility for vLLM and Hugging Face Transformers—demonstrating a rare mix of systems-level engineering and algorithmic depth. Ambitious and initiative-driven, he routinely translates research insights into practical, impactful deployments for industry and public sector clients.
7 years of coding experience
7 years of employment as a software developer
Master of Science (MSc) Mathematics, Master of Science (MSc) Mathematics at Københavns Universitet - University of CopenhagenMathematics and Philosophy, Mathematics and Philosophy at University of LeedsHigher Commercial Examination Programme Business/Commerce General, Higher Commercial Examination Programme Business/Commerce General at Niels BrockDoctor of Philosophy - PhD Mathematics, Doctor of Philosophy - PhD Mathematics at University of Bristol
Danish, English, German
Stackoverflow
174reputation
3kreached
1answer
3questions
Github Skills (13)
transformers10
lm10
regex10
python10
generative-ai10
llm10
prompt-engineering9
json9
keras6
multiprocessing6
neo4j6
cypher6
julia6
Programming languages (13)
Github contributions (5)
dottxt-ai/outlines
Jan 2024 - Feb 2025
Structured Text Generation
Role in this project:
ML Engineer
ML EngineerContributions:2 reviews, 6 PRs, 28 comments in 1 year 1 month
Contributions summary:Dan primarily focused on integrating and improving Outlines' compatibility with different language models, specifically vLLM and Hugging Face Transformers. Their contributions include fixing integration issues, such as pickleability problems with vLLM's `SamplingParams` and adapting tokenizers for compatibility. The user also implemented support for Transformers-based models via logits processors and extended regex functionality. They addressed issues related to the FSM logic, by returning the final state if already in the final state.
cfggenerative-aijsonllmsprompt-engineering
alexandrainst/NordjyllandNews
Jan 2023 - Dec 2024
Dataset containing news from Northern Jutland in Denmark.
Contributions:25 reviews, 5 PRs, 3 pushes in 1 year 10 months
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.