New York, New York, United States
Senior engineer with 10+ years of experience designing, building, and deploying large-scale AI systems and infrastructure at companies including Meta, Google, and Amazon.
- Designed and engineered AI products for businesses using the modern GenAI stack, including a multi-modal RAG search system, LLM-powered video insights tool, and agentic systems. - Led end-to-end engineering: from requirements gathering, through architecture design and building, as well as deployment to production and systems maintenance.
Received highest performance evaluation "Redefines Expectations" (top ~1% of company). News Feed Team: - Engineered AI models for text, image, and video understanding, and utilized multimodal and sparse neural network architectures. This involved end-to-end ML engineering, such as dataset engineering (handling PBs of News Feed data), model training, as well as model productionization. Reduced harmful content/misinformation distribution, and improved user sentiment towards the posts they see on News Feed. - Designed AI systems for prefetch targeting, reducing energy usage and saving millions of dollars in costs - Led various multidisciplinary workstreams (10+ personnel), leading engineers (both junior/senior), research scientists (AI PhDs), PMs, data scientists, designers, and UXRs. Neural Interfaces Team: - Engineered on Meta’s AR wristband, enabling a new way to interact with computers and making them easier to use for amputees - Migrated start-up acquisition CTRL-Labs’ modelling stack to Meta’s AI infrastructure
Improved throughput of various Google Maps features by +45%, by designing new data creation tools. Any time you see a curved (e.g. road bend) or circular (e.g. roundabout) feature on Google Maps, it was drawn using those tools.
Trained AI models to improve Amazon Search. Built hyper-parameter search system for AI models. Improved Amazon search model performances across the company, through dataset engineering work.