Senior Technical Program Manager | AI Builder | NVIDIA Inception Member
I build AI systems on hardware I own. 13 years of program management, 10 LLMs published on HuggingFace, and a DGX Spark running 24/7 on my desk.
GPU-native knowledge intelligence platform built on 6 NVIDIA technologies.
Your experts. Your GPU. Your data never leaves.
Ingest domain expertise at scale, build GPU-accelerated relationship graphs with RAPIDS cuGraph, and serve answers through any OpenAI-compatible tool. Built on NIM, TensorRT-LLM, Triton, NeMo Guardrails, cuGraph, and CUDA.
| Project | What It Does | Stack |
|---|---|---|
| NeuralForge | GPU-native knowledge intelligence with temporal knowledge graphs | NIM, TensorRT-LLM, Triton, NeMo Guardrails, cuGraph |
| CMMC Compliance AI | 10 fine-tuned LLMs for cybersecurity compliance (CMMC, NIST, HIPAA) | QLoRA, GGUF, Ollama, DGX Spark |
| Governed LLM Gateway | Policy-as-code gateway with tamper-evident audit trails | FastAPI, SHA-256 hash chains, 103 tests |
| Agentic Evaluation Sandbox | Doer/Judge/Adversary/Observer framework for agent testing | Multi-agent orchestration |
| Self-Healing Agentic Workflows | Circuit breakers, fallback chains, auto-reroute for autonomous agents | Failure detection, recovery |
| garak Contributions | Adversarial probes for NVIDIA's LLM vulnerability scanner | Prompt injection, Unicode obfuscation |
- NVIDIA DGX Spark (GB10, 128GB unified memory) running daily
- Gemma 4 26B A4B for inference at 43 tok/s
- 486K+ knowledge chunks from 80+ AI/ML experts
- 10G office network connecting DGX Spark, NAS, and workstations