Agnik Banerjee
Software Engineer
Computer Science student at the University of Maryland minoring in Technology Entrepreneurship, currently a Software Engineer Intern at Tesla. I build AI agents, ML systems, and distributed infrastructure, with past work at AWS, Amazon Leo, and Booz Allen Hamilton and published research on AI in healthcare.
Career Timeline
Professional journey and experience timeline.
Projects
Personal and research projects I've built.
Live Wikipedia trend tracker streaming 400K+ daily edits through Go and Kafka to detect breaking news as it happens.
- Built a live Wikipedia trend tracker in Go for 2K+ users, streaming 400K+ daily edits via Kafka into per-page counters
- Served search on 50K+ pages at 200ms p95 (3× faster) by indexing by day so queries skip old data, and caching in Redis
- Detected breaking news by flagging sliding-window edit-rate spikes per page, then summarizing each surge with an LLM
A Stanford University research game for families with children ages 3-12. Uses AI to analyze children's behavioral expressions from home video to aid autism research.
- Shipped an on-device facial expression classifier to 10K+ Android users, so therapy-session video never leaves the phone
- Shrank the model 14× at a 1.12% accuracy cost by quantizing weights to 8-bit integers and pruning near-zero ones in TFLite
- Built a Flower federated learning pipeline that keeps raw data on-device and shares only model weights, raising accuracy by 8%
Skills
Key skills that define my professional identity.
Python
General-purpose programming language for scripting, automation, and ML.
Java
Object-oriented language for scalable backend and Android development.
TypeScript
Typed superset of JavaScript for scalable web applications.
JavaScript
Create interactive and dynamic web experiences.
React
Build interactive UIs with components and hooks.
Next.js
Full-stack React framework with SSR and static generation.
Contact
Let's connect and explore collaborations.