Director of Platform Engineering at ZeroEyes, building AI infrastructure on FedRAMP-compliant Kubernetes.
5+ years leading platform teams, 10+ years hands-on across systems and cloud-native infra. Currently extending that discipline to the AI platform layer — MCP servers, RAG systems, GPU inference, LLM evaluation, and the security models regulated industries need for agentic workloads.
- 🧠 MCP servers and RAG systems in production (see
arxiver↓) - 🎯 Dimension-routed LLM evaluation on local GPUs (see
nite-eval↓) - 🔊 GPU-backed inference serving patterns (see
transcriber↓) - 🔐 FedRAMP Moderate platform buildout at work — AWS, EKS, Istio/Envoy, FIPS-140
- arxiver — Personal arXiv research assistant. FastMCP server, ChromaDB semantic search, TensorFlow recommendations, LLM summaries. Same backend exposed as CLI, REST API, Streamlit UI, and MCP.
- nite-eval — Autonomous overnight LLM evaluation pipeline. 15 multi-turn agentic tasks, dimension-routed dual-judge scoring, SQLite checkpoint/resume. Targets and judges on separate GPUs.
- transcriber — Streaming STT with NVIDIA Canary-1B on CUDA. FastAPI inference endpoint, chunked audio streaming, proof-of-pattern for GPU model serving.
- artemisee — Real-time 3D visualization of NASA's Artemis II using JPL Horizons ephemeris data. Hermite spline interpolation, client-side celestial math, 10 live data feeds.
- 🌐 woojay.com — AI-powered portfolio. Paste a job description, get a fit verdict.
- 📧 tenaciouswp@gmail.com
Python Go Kubernetes AWS GCP Pulumi OpenTelemetry llama.cpp MCP FastAPI Supabase Chroma Claude Code
AWS (Systems Development Engineer) · Apex Clearing (SRE) · LINBIT (software engineer, DRBD) · co-founded an embedded hardware startup · US Air Force officer