Twenty years deep in enterprise infrastructure — and I'm not slowing down. I build transformers from the attention layer up, design two-phase cooling for GPU-dense racks, and run production systems where real money is on the line. Two decades earned the depth. Staying current earns the trust.
Primary technical closer on the company's largest deal — 300 MW of AI-factory cooling. Promoted within five months. Built twenty-plus technical assets and thirty-plus proposals, reducing average sales cycle by roughly a third. Presented from engineering team to C-suite boardroom.
Architected scalable, cost-effective, HA solutions on Google Cloud Platform. Led infrastructure assessments and delivered prescriptive roadmaps that accelerated cloud adoption for mid-market and enterprise accounts. Mentored junior engineers on sales strategy and POC execution.
US West Region technical lead and primary closer. Drove adoption of hyper-converged infrastructure. Spearheaded Kubernetes and cloud-native mentorship across the engineering organization. Executed high-stakes POCs with custom automation.
Senior SE and architect roles across the major enterprise storage platforms — all-flash arrays, XIV, Symmetrix, CLARiiON. Managing Director of Systems Engineering at CreekPath. Technical Product Manager at Veritas. 127% quota · $10M+ territory.
GPU-dense data-center design, two-phase direct-to-chip cooling, and the power / thermal trade-offs that define AI-factory economics.
Stanford ML specialization. Hands-on with LLM internals — transformers from scratch, embeddings, attention mechanics. Building it, not just talking about it.
Triple AWS-certified. Production experience across GCP and AWS with Kubernetes, serverless, and hybrid architectures for enterprise workloads.
Twenty years across every major platform — EMC, NetApp, IBM XIV, Cisco HyperFlex. From Symmetrix frames to all-flash arrays to hyper-converged.
The rare engineer who runs a POC on Monday and presents to a CTO on Tuesday. Challenger methodology with a track record of outsized quota attainment.
Production code, not just slides. Custom tooling for sales automation, data-protection workflows, and AI-assisted proposals.
Arbitrage scanner, passive market-making with microprice fair-value and quarter-Kelly sizing, and weather contracts driven by calibrated NWS forecast models. Cross-market price discovery against Polymarket. Runs 24/7 under launchd with automatic settlement, take-profit, and reinvestment logic.
Working through Manning's Build an AI Agent (From Scratch) — building agents the hard way, no LangChain, no AutoGen, no framework abstractions hiding the mechanics. Tool-use loops, planning, memory, evaluation. Same philosophy as the LLM build: understand every layer from when to use to data pipeline design.
A transformer-based language model built from the ground up in PyTorch — tokenizer, attention, training loop. Understanding every layer of the stack: Embedding, Positional Encoding, Multi-Head Attention, Feed-Forward, Normalization, Linear Mapping, and Softmax output.