Journey
Work, study, and building — in rough reverse-chronological order.
Site Reliability Engineer, ML Platforms
Apple
Built a Kubernetes capacity forensics platform end-to-end (collectors, scanners, delta-query UI) used by 30+ SREs to diagnose EC2 capacity exhaustion, saving up to $3M per AWS availability zone annually.
MEng Computing (AI & ML)
Imperial College London
Studying AI/ML, systems programming, and compilers. Coursework spans distributed systems, functional programming, and low-level C. Expected graduation July 2027.
Full-stack & AI/ML Developer
8x
Optimized production analytics from 24s to sub-second via SQL-side aggregation and indexed Postgres RPC rewrites; cut /posts payloads 90%+ (22MB → ~1–2MB).
General Engineer
Canopy Labs
Architected, built, and deployed the company's flagship full-stack web application, enabling real-time multi-user usage with Docker and Kubernetes orchestration.
Full-stack & AI/ML Developer
Vani
Architected, shipped, and deployed the flagship multi-tenant web app on AWS with Docker + Kubernetes.
Machine Learning Developer
Trajex
Deployed LLama 3.2-7B-Instruct in production — 20% cost reduction vs OpenAI, 12% lower inference latency.
SwyftGesture — 1st Place
Hackathon
Hands-free computer control via webcam — mouse, scroll, and volume gestures using MediaPipe and OpenCV. Won 1st place against 40+ teams.
ML Engineer (Contract)
Altus Reach
Team of 3 — built a video saliency model improving prediction accuracy by 19%.
FRC Team Leader
#8235 Beyond the Flames
Led 20-student team for the FIRST Robotics Competition. Designed robots with Fusion 360, wrote autonomous control in Java and Python. Earned Dean's List nomination.