Case Studies
Production engineering work, written up plainly. Architecture decisions, agentic AI patterns, and the trade-offs behind the systems - not marketing.
Cost-Optimized Intelligent Document Extraction
A regulated income-verification IDP pipeline re-architected from managed frontier APIs to self-hosted Qwen open-weight models on vLLM, with open-source OCR and a confidence-gated model cascade.
The cost call, grammar-constrained JSON extraction, an OpenCV quality gate, managed frontier retained as a governed escalation, and an evaluation harness that blocks any cost win that regresses quality.
Read the case study → Case StudyAgentic Supply Chain Intelligence Platform
Greenfield agentic supply chain intelligence platform on Python FastAPI and vLLM with Qwen3.6-27B-class open-weight models, combining contract performance with utilization analytics. In production.
Architecture, two-pass contract extraction with per-field confidence scoring, AWS Textract OCR fallback, fuzzy plus semantic vendor matching, and a compliance scoring engine producing Critical/High/Medium/Low risk classification.
Read the case study →More case studies are added as engagements close. For the thinking behind these builds rather than the architecture, see insights, or get in touch to discuss specific engagements.
Looking for a senior engineering leader who can ship AI in production?
Open to a full-time senior engineering leadership role. No AI theater. Measurable outcomes only.
Hire Me →