Computer Science (Software Engineering) @ Arizona State University Β· 4.0 GPA Β· Graduating May 2027 π Tempe, AZ Β· π ayushwattamwar.com
I build scalable, intelligent software across the stack, from Rust systems and AI-integrated APIs to full-stack web apps and native iOS. I care about clean architecture, real performance numbers, and shipping things people actually use.
Halley β Your production traffic is your test suite.
An open-source, self-hostable LLM observability and regression testing platform. It records production AI agent runs as bit-fidelity cassettes, turns any run into a permanent regression test, and replays your whole fixture library in CI at $0 LLM cost, then bisects to the exact commit that broke it.
- βοΈ Rust ingester (tokio, axum, tonic), OpenTelemetry-native, normalizes 5 trace dialects into one canonical schema
- π Next.js 14 dashboard Β· TypeScript & Python SDKs Β·
halley ciCLI for replay and bisect - ποΈ Tiered storage: ClickHouse + PostgreSQL + Redis Streams
- π 4,792 spans/sec sustained Β· 0% error rate Β· 1.44M spans ingested with zero data loss
- π³ Docker Compose local Β· Kubernetes Helm for production Β· MIT licensed
- Halley β LLM observability & regression testing platform Β· Rust, Next.js, Python, Docker, Kubernetes
- Inference-time LLM Reasoning Agent β Multi-agent reasoning system; lifted accuracy from 31% to 38% over 6,500+ questions Β· Python, NLP
- Meal Planner & Recipe Finder β Native iOS app with offline sync and accessibility support Β· Swift, SwiftUI, SwiftData
- Academic Review System β Role-based MVC platform with 90%+ test coverage Β· Java, JavaFX, JUnit
- YouDub β Real-time YouTube voice-overs in 100+ languages, built at CalHacks Β· JavaScript, Google Cloud, ElevenLabs
Always open to internships, open-source collaboration, and good tech conversations.