Sayak Paul
Senior Research Engineer at Hugging Face
Kolkata, West Bengal, India
Summary
🤩
Rockstar🎓
Top SchoolSayak Paul is a Senior Research Engineer at Hugging Face with nine years of experience building and shipping state-of-the-art image and video generation systems, primarily via the diffusers library. He blends research and engineering—training diffusion models, contributing core features, performance work, and end-to-end examples—while shepherding kernels and collaboration projects like SPRIGHT and ReflectionFlow. An active open-source contributor, he has ported and implemented major vision models in Transformers and Keras (ConvNeXt, ViT MAE, Data2VecVision, SegFormer) and authored practical notebooks and LoRA fine-tuning recipes used widely by practitioners. Previously he built scalable ML pipelines and MLOps at startups, improving production latency and reproducibility, and authored tutorials and docs that helped teams adopt TPU and cloud workflows. Based in Kolkata, he prefers technically rigorous conversations and is selective about recruiter outreach, reflecting a focus on deep technical fit. Outside of release cycles he quietly balances model hacking with clear, research-driven documentation—a combination that surfaces in both his code contributions and educational writing.
9 years of coding experience
9 years of employment as a software developer
High School, Science (Physics, Chemistry, Maths, Computer Science), High School, Science (Physics, Chemistry, Maths, Computer Science) at Jadavpur VidyapithBachelor of Technology - BTech, Information Technology, Bachelor of Technology - BTech, Information Technology at Netaji Subhash Engineering College
Hindi, Bengali, English
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (34)
transformers10
pytorch10
artificial-neural-networks10
python10
classification10
image-classification10
model-driven10
configuration-management10
machine-learning10
stable-diffusion10
classify10
model-building10
keras10
deep-learning10
trainings10
Programming languages (17)
Github contributions (5)
huggingface/diffusers
Jan 2023 - 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:20 releases, 5524 reviews, 11 commits in 9 days
Contributions summary:Sayak was primarily involved in adding examples and fine-tuning scripts related to LoRA models within the Hugging Face Diffusers repository. Their work focused on fine-tuning the Stable Diffusion XL (SDXL) models for image generation tasks using the DreamBooth and InstructPix2Pix techniques. The user's commits demonstrate expertise in integrating and utilizing LoRA (Low-Rank Adaptation) techniques for model adaptation, specifically for text-to-image tasks, suggesting a specialization in machine learning engineering.
pytorchartdeep-learningimage2imagestate-of-the-art
keras-team/keras-io
May 2020 - Jan 2023
Keras documentation, hosted live at keras.io
Role in this project:Technical Writer
Technical WriterContributions:189 reviews, 42 commits, 42 PRs in 2 years 8 months
Contributions summary:Sayak's commits primarily involve modifications to the `keras-io` repository's documentation. These changes include adding and updating documentation related to various machine learning topics, such as distributed training, image augmentation, and model fine-tuning. The contributions mainly focus on providing clearer explanations, examples, and code snippets to help users understand and utilize the Keras framework. The user has improved the clarity and completeness of several existing documentation sections, specifically relating to topics in image and video processing.
deep-learninghostedmachine-learningtensorflowkeras
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.