About
Senior software engineer with a background in semiconductor memory design — SRAM arrays, memory compilers, and advanced process nodes. That foundation shapes how I think about every layer of the stack above it.
For the past year I’ve been building production AI systems: RAG pipelines, LangGraph agents, inference tooling. Now going deeper into the inference layer and systems architecture — the engineering that makes AI fast and viable at scale.
This is my working log. Real experiments, real numbers.
Currently building
- LLM inference benchmarking suite
- AI solutions architecture portfolio
- Triton kernel experiments
Writing about
- Inference engineering — KV cache, quantization, batching, vLLM internals
- GPU architecture and performance profiling
- Enterprise AI architecture patterns
- Project retrospectives with benchmark numbers