Case studies

Case studies in production-aware AI systems.

Scan by domain, then open the systems with enough public evidence to evaluate architecture, implementation choices, and responsible-use boundaries.

Healthcare & Insurance AI

Standards-aware APIs, event-driven healthcare platforms, quality workflows, and claim/document automation.

FHIR response contractHL7 architecture diagramsOCR extraction JSON
Lead case study

Hospital Readmission FHIR ML API

Synthetic FHIR-style scoring service for 30-day readmission risk with schema validation, feature extraction, explainability, tests, model-card framing, and Dockerized FastAPI deployment.

flagship
Open case study
Healthcare interoperability / quality intelligenceHL7 AI Challenge

Public challenge system with synthetic/sample boundaries

Computer vision / OCR / structured extractionFreshTrack AI

Production-style reference implementation

Agentic AI / LLMOps

Local model workbenches, agent state, tool calls, verification gates, and experiment surfaces.

activation steering GIFhook placement diagramagent verification path
Lead case study

LLM Steering Lab

Local-first activation steering and representation-engineering lab with ActAdd-style vector extraction, FastAPI workbench APIs, React UI controls, model support notes, and reproducible artifacts.

flagship
Open case study
Agentic market intelligence / quant researchAlphaQuant

Engineering demo; not investment advice

Document AI / RAG

Document parsing, OCR, RAG, prompt construction, local inference, confidence scoring, and review surfaces.

local inference boundaryreceipt extraction samplePrompt Flow RAG graph
Lead case study

Local Document AI Extraction

Private/local PDF extraction pattern using PDF parsing, chunking, LangChain, Streamlit, Dockerized Ollama, analyst-authored criteria, and reviewable outputs.

polished demo
Open case study
Computer vision / OCR / structured extractionFreshTrack AI

Production-style reference implementation

Azure ML / prompt-flow RAGChatWithWiki AzureML

Prompt Flow RAG demo

Quant & Forecasting

Forecasting and research systems framed around baselines, calibration, evidence, and responsible use.

Fusion report shape2024 Brier tablehistorical backtest JSON
Lead case study

Agentic Alpha Engine

Local-first agentic market intelligence workbench with public-signal ingestion, retrieval/storage services, planner agents, verification rules, structured reports, and Docker architecture.

flagship
Open case study
Sports forecasting / model calibrationNFL Forecasting Lab

Historical forecasting lab; not betting advice

Research Archive

Earlier healthcare, RAG, visualization, infrastructure, and ML foundation projects that show technical range and learning depth.

LOS regression notesHEDIS RAG conceptGPU planning matrix
Lead case study

Azure LOS Prediction

Length-of-stay regression work using a Microsoft toy hospital dataset, Azure Notebooks, and baseline regression methods.

experiment
Open source evidence
Healthcare quality / RAG conceptHEDIS RAG SSD

Concept prototype; public-safe framing required

AI infrastructure planningGPU Analysis

Infrastructure thought-work

Deep learning visualization / researchHNSC Viz

Research visualization note

Privacy boundary

Public-safe by design.

All public examples are synthetic, sanitized, historical, or research/demo implementations. No PHI, PII, employer-confidential data, proprietary claims data, private keys, production credentials, or sensitive datasets are included.