Machine Learning Engineer at Meta working on ads auction optimization. Previously built LLM infrastructure at SliceX AI, researched efficient attention mechanisms at Microsoft, and shipped front-end features at Google and Amazon.
M.S. in Artificial Intelligence from Northwestern University.
ML & AI -- by day I work on large-scale ML models for ad systems. Outside of work I stay hands-on with deep learning fundamentals, from implementing transformers from scratch to experimenting with local LLM deployments on my homelab.
Self-hosting -- I run a Proxmox-based homelab with reverse proxies, VPN mesh networking, containerized services, and local AI model serving. At big companies, most infrastructure is abstracted away -- running my own forces me to actually understand networking, DNS, TLS, storage, and deployment at a level you don't get when everything is a managed service. It makes me a better engineer at work.
Track days & data -- I drive a Subaru BRZ on track and combine my engineering mindset with motorsports. I built 86 Challenge -- a telemetry web app for the GR86/BRZ/FRS time trial community where drivers can upload RaceChrono data, compare laps, and track progress. I also write about my track experiences and car modifications on my blog.
- 86 Challenge -- Telemetry app for the 86 time trial community
- Blog -- Track days, car mods, and other write-ups