AI/ML Systems Engineer
Building production-grade ML infrastructure — from custom GPU kernels to agentic architectures.
| Project | Description | Stack |
|---|---|---|
| Rusty | GPU-accelerated ML framework with custom compute kernels, LoRA fine-tuning, and Llama inference | Rust, wgpu, Metal, WGSL |
| ARA | Advanced Reasoning Agent — ReAct loop with self-reflection and tool use | Python, LangGraph |
| Archi | Website reverse-engineering for architect-grade implementation prompts | Python, Playwright |
ML Infrastructure: Rust, WGSL, wgpu, Metal, safetensors, GGUF
Deep Learning: JAX, Flax, PyTorch, Transformers
Agents: LangGraph, OpenRouter, E2B
Interpretability: TransformerLens, SAELens
- Custom GPU kernels and ML compute backends
- LLM inference optimization and quantization
- Agentic systems with structured reasoning
- Mechanistic interpretability