Computer Scientist 路 Applied AI, cloud & on-premise
I build enterprise AI systems for large clients, nationally and internationally.
Legal document retrieval I build retrieval systems over legislation and case law. They ingest new documents daily, detect duplicates using each document's own official identifier, and route queries across several indexes so that the source and the court type stay independent of each other. Most of the engineering goes into keeping that pipeline reliable rather than into choosing an embedding model.
Agent workflows over enterprise data Intent routing into specialized agents, shared conversation memory, and tool calling against live business systems including SAP and Azure SQL.
Access control and encryption for AI products Per-owner envelope encryption for chat and document content, tenant isolation driven off directory claims, and a single authorization choke point with no admin bypass.
Cost control Per-user token budgets metered at the completion funnel rather than per route, so adding an endpoint can't silently route around the meter.
Computer vision YOLO and ResNet architectures for industrial inspection.
| Cloud | Azure (AI Foundry, OpenAI, AI Search, Functions, Cosmos DB, Entra ID) 路 AWS 路 Google Cloud |
| On-premise | Self-hosted and locally deployed AI services, for data-residency and air-gapped requirements |
| Certified | Microsoft Certified Trainer 路 Azure Solutions Architect Expert 路 Power Platform Solution Architect Expert |
I draw, model in 3D, and build games. It isn't a separate hobby from the engineering, it's where the interface instincts come from.
Note
Most of the AI platform work lives in private client repositories.