Looking for a job

John Pritchard

@jhpritch-dev

Message

AI Infrastructure Engineer building privacy-first distributed LLM inference and platform-grade MLOps systems.

United States

Message

What I'm looking for

I’m looking for Remote AI Infrastructure, Platform Engineering, MLOps, or SRE roles—especially in compliance-heavy industries (healthcare, legal, finance, federal, insurance) where privacy-first AI is required, not preferred.

I build the infrastructure layer behind distributed AI inference, so LangGraph can decide which model runs, why, and what to do if it fails. My flagship BIFROST RAG system is a production-grade, distributed inference platform running on three nodes with Kubernetes orchestration, eight inference endpoints, and a custom German (Elementary) complexity classifier that routes each request in real time—delivering 86% cost reduction versus cloud-only AI without sacrificing capability.

I also create autonomous, verifiable pipelines like AUTOPILOT (LangGraph-based), where tasks are decomposed into parallel DAGs and executed with hierarchical escalation across NPU/iGPU/GPU tiers, then validated through a six-level objective harness before returning results. With JARVIS-OFFLINE mode, I enable zero-cloud-egress operation for air-gapped or compliance-restricted environments, backed by production observability (Prometheus/Grafana), zero-trust networking (Tailscale/WireGuard), and a modern platform stack (K3s/Kubernetes, FluxCD GitOps, ChromaDB/Qdrant, ROCm GPU optimization, and Pydantic v2 contracts).

Experience

Work history, roles, and key accomplishments

Current

AI Infrastructure Engineer

Current

Independent

Jan 2019 - Present (7 years 5 months)

Designed and operated BIFROST, a 3-node distributed AI inference platform with 8 endpoints and Kubernetes orchestration, reducing AI inference costs by 86% versus cloud-only with no capability loss. Built LangGraph-based AUTOPILOT task routing with tiered NPU/iGPU/GPU escalation and enabled JARVIS-OFFLINE zero-cloud-egress deployments with production observability and zero-trust networking.

Kubernetes LangGraph ROCm GPU Optimization Zero Trust Networking (Tailscale WireGuard)RAG (ChromaDB Qdrant)fastAPI Python

Embedded Systems Engineer

Henry Hawkinson

Oct 1990 - Jul 2023 (32 years 9 months)

Installed and maintained CPU-driven electronic control systems, relay logic, and electromechanical integrations, performing systematic diagnosis across digital control, analog circuits, and physical components. Supported client maintenance agreements across South Carolina and delivered complex hardware restoration and microprocessor switching system installations.

Embedded Systems Circuit Troubleshooting Digital Control System Relay Embedded Maintenance

Adjunct Music Professor

Christopher Newport University

Aug 2021 - May 2023 (1 year 9 months)

Taught foreign language diction (German, French) and aural skills, delivering instruction across multiple instructional modalities. Administered online course delivery platforms and digital assessment tools to support course evaluation and student progress.

German Diction Instruction Aural Skills Teaching

Assistant Store Manager

uBreakiFix

Mar 2018 - Aug 2018 (5 months)

Provided PC and mobile device hardware/software diagnostics and repair, delivering customer-facing technical support across varied technical literacy levels. Trained junior technicians on systematic troubleshooting methodology.

Technical Troubleshooting Customer Technical Support Technical Training

Doctoral Assistant

University of South Carolina

Aug 2010 - May 2013 (2 years 9 months)

Designed an ad hoc WiFi wireless telemetry system for real-time physiological data collection and deployed it using IoT sensors and wireless networking. Presented system architecture to a research committee and supported practical IoT infrastructure deployment in a high-stakes environment.

Wireless Telemetry System Design IoT Sensor Deployment Networking Signal Processing IoT Infrastructure Deployment

Assistant Professor

Kennesaw State University

Aug 2002 - May 2006 (3 years 9 months)

Developed online course delivery platforms across online-only, hybrid, and traditional modalities. Implemented live student-response systems for real-time assessment of 100+ students and built PC-based electronic assessment with automated grade integration.

Online Course Delivery Platforms Hybrid Instruction Real Time Large Group Assessment Live Student Response Systems Automated Grade Integration