Case Studies
Real results from real businesses.
What this looks like in practice
A few stories from teams that recruited AI — and won.
Capacity Planning Model
A team at LinkedIn faced a 9-figure hardware allocation problem in their proprietary data centers, manually forecasting capacity with spreadsheets every quarter.
I built them a model that reasons over their historical data and constraints — so they could ask questions in plain English instead of wrestling with formulas.
13.4% improvement over spreadsheet heuristic in predicting hardware provisioning needs, resulting in substantial cost savings.
Research Acceleration
ML workloads at AMD were hitting a wall — too slow, too expensive, too power-hungry.
I designed a compute strategy that paired the right models with the right hardware, including some analog AI tricks most teams don't know exist.
3–5x faster runs. Lower power bills. Same accuracy.
Knowledge Retrieval System
Critical know-how at TechnicallyFit was scattered across wikis, Slack threads, and people's heads.
I deployed a RAG system over their internal docs — so anyone could search and get real answers, not just links.
New hires ramped 40% faster. Less time reinventing the wheel.