I'm a petroleum engineer who built an AI enablement platform at a major E&P operator and trained 300+ people on Claude Code - from individual contributors through the leadership team, including three C-suite members. Trained engineers have self-reported $12M+ in savings. One prevented a loss of primary containment event. I teach the same workflows externally.
Petroleum engineer by training. I've spent 14 years across production engineering, SCADA systems, and enterprise software - at Chevron, Xcel Energy, and a major upstream E&P operator. I build tools that engineers actually use.
B.S. Petroleum Engineering from Colorado School of Mines. Started in economic modeling and well abandonment at Chevron, moved through SCADA systems at Xcel Energy (gas pipelines and electric transmission), and landed in production systems engineering where I build data-driven tools, ML systems, and cloud infrastructure for upstream operations.
In early 2026 I built a company-wide AI enablement platform from scratch: MCP infrastructure for database access, a provisioning portal, a self-service Heroku-style deployment platform, and a training program that put AI coding tools in front of 300+ engineers. All organic demand.
Practical, hands-on training for engineers who want to use AI coding tools on real engineering problems. Not a lecture. Not "intro to AI." You bring your data, we solve your problems live.
Most AI training is taught by people who have never run a decline curve, read a SCADA alarm, or argued with a vendor about polling rates. I have. The workflows I teach are the same ones that produced $12M+ in documented savings at a major E&P operator - adapted for your team's data, your systems, and your engineering problems.
First pilot cohort is in the works. To get notified when registration opens, reach out at the email below or connect on LinkedIn.
Enterprise AI platforms on the clock. ML, embedded systems, and precision timing off it. The point is the same either way: I build things that work, end to end.
Built solo in months: MCP infrastructure connecting AI tools to company databases, a credential provisioning portal, and a self-service platform that security-scans, containerizes, and deploys engineer-built apps to Kubernetes - no IT tickets. Dozens of production apps shipped by engineers, not developers.
Multi-model deliberation system: routes a prompt to 11 Bedrock models concurrently, shuffles responses to eliminate position bias before voting, and synthesizes a final answer with a chairman model.
AR plane spotting app. Custom YOLOX and D-FINE detection models trained from scratch, SwiftUI + CoreML on-device inference, C# WebSocket broadcast service with a physics engine, Django/PostGIS backend, live ADS-B data pipeline.
Precision NTP timing on Raspberry Pi with GPS PPS. Thermal management, stability analysis, and the engineering behind sub-microsecond accuracy.
PID-controlled autopilot for X-Plane built in Python. Real-time control loop via Flask, Redis, and WebSockets. Full-stack systems engineering.
Reverse-engineered the SMBus protocol for Inventus 24V LiFePO4 batteries with no public documentation. Full register map, per-cell voltages, health profiling across a 5-pack battery bank.
More at austinsnerdythings.com - ML/CV, DevOps, embedded systems, IoT, SCADA, storage engineering, and network architecture. This site is served from a colocated Dell R630 I run for $55/month (equivalent AWS bill: ~$2,000).
Interested in training for your team? Have a question? Reach out directly.
[email protected]