About Me
Machine Learning Engineering
Builds production-grade ML solutions for resource-constrained and mission-critical environments. Leverages a GCP ML pipeline and expertise in graph neural networks and edge computing — delivering customer-validated systems through rapid iteration and real-world test events.
Defense Sector Expertise
15+ years of military operations experience inform reliable, deterministic AI for disconnected and high-stakes settings. Proven ability to translate complex battlefield needs into scalable technical solutions.
Leadership & Communication
Led teams of 5–30+ in tactical and strategic roles, briefing senior leaders and driving decisions under pressure. Skilled at aligning technical teams with mission objectives in dynamic environments.
Projects
What I'm building now — from local-first AI infrastructure to shipped apps
mnemos
Air-gappable, self-hosted memory layer for AI conversations. Phase 0 ships GPU inference on a single-node k3s cluster: NVIDIA device plugin, llama.cpp CUDA server, and an OpenAI-compatible endpoint.
Drone Object Detector
Object detection model trained on drone footage using YOLO architecture. Processes video clips to identify and track objects in aerial imagery.
Latest Writing
Engineering notes from projects I'm building in the open
Building mnemos
Phase 1: Platform Foundation
3 min read
Phase 1 of mnemos is done: Helm charts, Argo CD GitOps, and Prometheus/Grafana on the same k3s GPU node. Nothing gets applied by hand again — and the cluster is visible from a browser.
Building mnemos
Phase 0: GPU Inference on Kubernetes
4 min read
The first phase of mnemos, an air-gappable memory layer for AI conversations, is done: one k3s node, one GPU, one OpenAI-compatible endpoint. Here is what shipped and what comes next in the series.