Co-Founder & CTO of Krewva — the AI for your relationships. It reads across the apps you already use (iMessage, WhatsApp, Gmail, Calendar, Slack), surfaces the obligations hiding in your conversations, drafts the action in your voice, and acts only with your approval.
Math-stat background, self-taught into ML systems. I care about AI that's trustworthy in the real world, not just the demo.
Previously founded Machina AI — zero-cloud visual inspection for high-mix manufacturing. Built EdgeRunner, signed a pilot LOI with a Shenzhen powder metallurgy manufacturer, and learned that ToB sales on a single-founder timeline is a different problem than the engineering.
I also open-source SMBoost — a reliability harness for 2B–7B local LLMs, born from a tail-behavior failure mode I hit running a 3B VLM on Jetson during the Machina pilot. That same obsession — making small models trustworthy in production — is what Krewva runs on.
→ krewva — The AI for your relationships. Interpretation → decision → action, across your apps.
→ edge-runner — Real-time vision system on NVIDIA Jetson. YOLO + VLM dual-thread architecture. <12ms P99 latency.
→ smboost — Reliability harness for 2B–7B local LLMs. State machine + GBNF constraints + adaptive scoring.
Python · NVIDIA Jetson · TensorRT · YOLO · VLMs · LangGraph · llama.cpp · Computer Vision · ML Systems · Edge Deployment
w.haiyang@outlook.com · haiyang5535.github.io · linkedin.com/in/h-wu



