Aleksander Ficek
Senior Research Scientist at NVIDIA
London, England, United Kingdom
Summary
🤩
Rockstar🎓
Top SchoolAleksander Ficek is a Senior Research Scientist based in London with 8 years of experience building and deploying large-scale ML and LLM systems at NVIDIA, where he led work on Nemotron reasoning models, hybrid Nemotron-H families, and state-of-the-art supervised fine-tuning and evaluation pipelines. He combines deep research credentials (papers across NeMo, LLM reasoning, and synthetic verifier generation) with hands-on engineering—shipping GPU-accelerated algorithms (k-NN improvements in cuML) and production microservices for LLM evaluation and inference. Comfortable across Python, C++/CUDA, and infrastructure, he has driven RAG and PEFT projects accepted to EMNLP and helped transition research into Early Access products and a patent submission. A former visiting researcher at Cambridge and Berkeley, Aleksander blends academic rigor with practical system-building, and outside of work he ran ZurichAI and actively documents his output on Google Scholar and GitHub.
8 years of coding experience
6 years of employment as a software developer
Bachelor of Applied Science - BASc, Mechatronics Engineering, Bachelor of Applied Science - BASc, Mechatronics Engineering at University of Waterloo
Stackoverflow
35reputation
4kreached
0answers
6questions
Github Skills (22)
python10
nearest-neighbors10
machine-learning10
machine-learning-algorithms10
gpu10
cuda10
cuml10
testing9
c-language9
pytest9
cython9
cprogramming-language9
scikit8
scikit-learn8
dispatch6
Programming languages (10)
Github contributions (5)
rapidsai/cuml
Jul 2020 - Dec 2020
cuML - RAPIDS Machine Learning Library
Role in this project:
ML Engineer
ML EngineerContributions:2 reviews, 32 commits, 2 PRs in 5 months
Contributions summary:Aleksander primarily focused on improving the nearest neighbors functionality within the cuML library, a RAPIDS Machine Learning Library. They implemented and refined the `kneighbors` function and its related tests, including adding functionality to allow the user to input their own kNN graph. Their work involved modifying Cython code and updating tests, demonstrating a focus on enhancing the library's capabilities related to GPU-accelerated machine learning algorithms, specifically k-NN.
cudacumlnvidiadata-sciencegpu
aleksficek/Solar
May 2019 - Jan 2020
IoT 'smart' solar panels with API's in Node.js and C++, a React front-end and MongoDB for data logging
Contributions:21 commits, 2 PRs, 17 pushes in 8 months
apireactnode-jsapi-ssolar
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.