// Projects

The
Work

All projects
01
AI · JavaScript

Zuna

A sweet book narrator that turns readable PDFs and text files into audio locally in the browser.

8WASM threads
02
Inference · Python

content-moderation.ai

A transformer inference pipeline with ONNX export, graph rewrites, int8 quantization and FastAPI serving.

<25mstarget inference latency
03
Developer tooling · Python

ares-cli

A local coding environment manager for Apple Silicon with a live terminal dashboard.

2managed model roles
04
Inference Engineering

inference-1

rewriting SmolLM2-135M logic from scratch and building an inference engine around it

2xfaster inference