San Francisco, California, United States
I'm a Software Engineer at Meta working on Live AI — Meta's real-time video assistant that ships on Ray-Ban Meta smart glasses. It lets users have an ongoing conversation with AI while showing it the world around them. Think of it as an AI companion that can see what you see, remember what you showed it, and help you in the moment — all through your glasses. I've contributed to core features including: → Visual Memory — so the assistant remembers what you showed it earlier → Text Understanding — reading posters, tickets, signs, and documents in real-time → Agentic AI — guiding users step-by-step through tasks like cooking, repairs, and DIY projects Across each feature, my work spans the full LLM post-training data lifecycle — researching and experimenting with methods to generate high-quality synthetic data, sourcing from live traffic and third-party datasets, curating SFT training datasets, building evaluation pipelines, designing evaluation frameworks and rubrics, and building LLM-as-Judge pipelines with frontier models to measure and ensure model quality before shipping. Before Meta, I was at Samsung Research America (STAR Labs), where I re-architected the ML pipeline for NEON — generative AI-powered virtual humans. The work focused on pipeline design and systems optimization: modularizing components to reduce dependencies, improving reproducibility through version tracking and provenance, and introducing multi-threading and multi-processing — cutting end-to-end operation time by 50%. Master's in Engineering Design from Carnegie Mellon. I'm passionate about building AI that genuinely helps people, solving hard problems, and ensuring the technology we create is ethical and human-centered. Always happy to connect — feel free to drop me a note with your request! 😊
Working on Live AI, Meta's real-time video assistant for Ray-Ban Meta smart glasses that lets users interact with AI while showing it the world around them. I have contributed to features like visual memory (so the AI remembers what you showed it), text understanding (reading signs, posters, documents), and agentic AI (guiding users step-by-step through tasks). Across each feature, I have built training pipelines, curated datasets, and evaluation systems to ensure quality.
NEON (STAR labs) is a Samsung Research Lab that is dedicated to creating realtime hyper-realistic generative AI-powered virtual Humans called NEONS who interact with you in ways that preserves human social interactions and can advance the digital world. https://neonlife.ai/ Working on ML pipeline within the research team
EY is a consulting firm where I worked with a financial investment company that offered investment and retirement plans and tools to the customers. My responsibilities: Backend development, API design, CI/CD pipeline, system design, production support, unit testing
Aspire360 aimed at connecting new startup founders with founders and CEOs of already successful startups as mentors to help them navigate the early steps and finances of running a startup and how to navigate the market. My responsibilities : Team Lead, Developed the frontend and backend of the web application that would connect these founders to the mentors and create a customizable coursework for them
NoRILLA is a patented mixed - reality educational system bridging physical and virtual world to improve STEM learning in kids. https://www.norilla.com/ My responsibilities: Integrating Intel realsense camera with the existing code, 3d segmentation, object detection, calibration and on-site debugging, research and development