Suraj Patil
Research Scientist at Black Forest Labs
Paris, Ile-de-France
Summary
🤩
Rockstar🎓
Top SchoolSuraj Patil is a research scientist and machine learning engineer with nine years of experience building and optimizing state-of-the-art generative models and datasets, most recently at Black Forest Labs after a multi-year tenure at Hugging Face. He has hands-on expertise in diffusion and transformer models—contributing to flagship open-source projects like diffusers, transformers, and the widely used dalle-mini—focusing on efficiency features such as gradient checkpointing, TPU support, and model integration across Flax and PyTorch. His work expanding the huggingface/datasets library reflects a pragmatic commitment to broadening access to diverse NLP corpora, while notebook and pipeline contributions show he cares about reproducible, user-friendly tooling. Based in Paris, he blends research rigor with production-savvy engineering and a history of startup founding, making him adept at moving cutting-edge models from experiments into real-world workflows.
9 years of coding experience
8 years of employment as a software developer
Bachelor's degree Computer Science, Bachelor's degree Computer Science at Vivekanand CollegeMasters Computer Science, Masters Computer Science at Shivaji University, Kolhapur
Github Skills (23)
transformers10
pytorch10
python10
attention-mechanism10
diffusion-models10
machine-learning10
stable-diffusion10
tpu10
diffusers10
flax10
datasets10
transformer-models10
deep-learning10
image-generation10
nlp10
Programming languages (6)
Github contributions (5)
huggingface/diffusers
Jun 2022 - Jan 2023
🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch and FLAX.
Role in this project:
ML Engineer
ML EngineerContributions:1 release, 688 reviews, 349 commits in 7 months
Contributions summary:Suraj appears to be working on a deep learning project related to diffusion models. Their commits involve implementing and modifying core components within the Hugging Face diffusers library, including dynamic module loading, UNet modifications (e.g., for GLIDE), and the Stable Diffusion XL inpainting pipeline. The code changes indicate a focus on model integration, optimization, and extending the capabilities of diffusion models.
pytorchartdeep-learningimage2imagestate-of-the-art
huggingface/transformers
May 2020 - Nov 2022
🤗 Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX.
Role in this project:ML Engineer
ML EngineerContributions:738 reviews, 203 commits, 485 PRs in 2 years 6 months
Contributions summary:Suraj's contributions focused on improving the functionality and performance of the model evaluation process. They addressed issues with loss calculation in evaluation, particularly on TPU devices, and optimized the use of global attention masks. The work involved modifying the trainer and model code, and adding unit tests, indicating a focus on model development and quality assurance within the project.
pythonbertspeech-recognitionstate-of-the-artflax
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.