Iulia-maria Comşa
Research Engineer at Google
Romania
Summary
🤩
Rockstar🎓
Top SchoolIulia-maria Comşa is a Research Scientist at Google DeepMind with a decade of experience bridging machine learning research and production-quality software engineering. She holds a PhD in Neuroscience from Cambridge and combines deep academic training in computational neuroscience with hands-on systems work from roles at Google and Microsoft. Her open-source contributions include core compression improvements to the JPEG XL reference implementation and experimental ML layer tuning within Google Research, showing fluency across image codecs and model development. At Google she progressed from intern to software engineer before transitioning to research, evidence of strong execution in both product and research environments. Based in Romania, she brings a rare mix of theoretical rigor and practical impact—optimizing low-level algorithms while shaping ML experiments. Colleagues can expect a researcher-engineer who moves fluidly between high-dimensional modeling and performance-focused engineering.
10 years of coding experience
1 year of employment as a software developer
Master of Research, Computational Neuroscience and Cognitive Robotics, Master of Research, Computational Neuroscience and Cognitive Robotics at The University of BirminghamDoctor of Philosophy (PhD), Neuroscience, Doctor of Philosophy (PhD), Neuroscience at University of CambridgeUniversitatea Babeș-Bolyai din Cluj-Napoca
Romanian, English, French, German, Italian, Japanese
Github Skills (19)
algorithm10
algorithms10
c-language10
jpeg-xl10
python10
machine-learning10
data-structure10
tensorflow10
data-structures10
cprogramming-language10
image-compression10
decode9
encoder9
mlops9
encode9
Programming languages (4)
Github contributions (5)
libjxl/libjxl
Jun 2021 - Sep 2021
JPEG XL image format reference implementation
Role in this project:
Back-end Developer
Back-end DeveloperContributions:14 reviews, 5 commits, 10 PRs in 3 months
Contributions summary:Iulia-maria primarily contributed to the core functionality of the JPEG XL image format reference implementation. Their work focused on enhancing the modular palette transform, including adding features to improve color selection and quantization. They added a function to find frequent color deltas and improved penalty heuristics for delta selection, improving image compression. The user also allowed for delta palettes in >8-bit images, expanding the supported use cases.
lossless-image-compressionjpegimagelossless-compression-algorithmpsd
google-research/google-research
Sep 2020 - Oct 2020
Google Research
Role in this project:ML Engineer
ML EngineerContributions:7 commits in 6 days
Contributions summary:Iulia-maria primarily contributed to experiments and model configurations within the `google-research/google-research` repository. Their work involved modifying the MLayer, a custom layer, and tuning hyperparameters for various machine-learning experiments, specifically on CIFAR-10 and periodic functions. The user also updated and refined the MLayer experiment notebooks, which included adjusting model architectures and training procedures. These changes indicate an involvement in model development and experimentation within the research context.
googlemachine-learningai
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.