Sudeep Agarwal is a product leader and hands-on builder with 11 years of experience shaping growth and product strategy, currently serving as Vice President of Product Management at Tubi in San Francisco. He combines enterprise and consulting roots (Oracle, McKinsey, Google) with startup-style execution, scaling user growth and monetization across streaming products. Technically fluent, he contributes code to open-source ML projects—adding front-end Chrome-extension UI and REST/Python modules for in-browser and deployed LLM inference in notable repos like mlc-ai/web-llm and mlc-ai/mlc-llm. That mix of product leadership and full-stack development lets him move quickly from customer insights to deployable ML features. He holds a BE (Hons.) from BITS Pilani and an MISM from Carnegie Mellon, reflecting a strong foundation in both engineering and management. Colleagues describe him as a pragmatic strategist who still enjoys shipping code that directly impacts users.
11 years of coding experience
14 years of employment as a software developer
BITS Pilani, Birla Institute of Technology and Science
Universal LLM Deployment Engine with ML Compilation
Role in this project:
Full-stack Developer
Contributions:26 reviews, 19 PRs, 1 push in 5 months
Contributions summary:Sudeep implemented significant features related to model loading, configuration, and deployment within the MLC-LLM framework. Their work included adding support for custom model paths, downloading weights from Hugging Face, and integrating various model configurations (Llama, GPT-NeoX, Moss). Furthermore, the user contributed to the development of a Python chat module and a REST API, enhancing the framework's usability and deployment capabilities. They also refactored the REST API and added support for the Gorilla model.
Contributions:1 review, 8 PRs, 7 comments in 5 months
Contributions summary:Sudeep primarily contributed to the front-end development of a Chrome extension for the `web-llm` project. They implemented the UI components for the extension, including the input field, submit button, loading indicators, and answer display. Additionally, the user integrated the `ChatRestModule` and `ChatModule` for interacting with both a local server and the WebGPU backend, facilitating user interaction with the model. They also added support for stuff QA using the context in the active tab.
chatgptdeep-learninglanguage-modelllmtvm
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.