01Representation engineering
LLM Steering
Tests, workbench UI, research records, checked-in outputs, demos
- Status
- Maintained · green CI · Pages
- Limit
- Research workbench; effects vary by model, layer, prompt, and coefficient.
PROJECT LAB / 39 AUTHORED REPOSITORIES AUDITED
Projects & experiments.
A collection of open-source systems and research experiments. Browse the code, working interfaces, and the questions behind each project.
Selected projects: LLM Steering · Meta Harness · Readmission FHIR ML API · HL7 AI Challenge
Browse the collection
Inside the projects.
Local AI, healthcare interoperability, agent workflows, and scientific computing.
Project explorerRepository snapshot · July 2026
Featured / 6
Selected implementations.
Code, architecture, and technical documentation, with each project's current scope.
02Agent systems
Meta Harness
Schemas, CLI/MCP, proof ledger, integrations, examples, docs
- Status
- Pre-1.0 · 14 green checks
- Limit
- Source-installed and pre-1.0; correctness is not guaranteed.
03Predictive healthcare
Hospital Readmission FHIR ML API
Synthetic FHIR, explainability, tests, docs, Docker
- Status
- Maintained · green CI
- Limit
- Deterministic demo model; not clinically validated.
04Healthcare interoperability
HL7 AI Challenge Platform
Architecture, demo package, event-driven services, test scripts
- Status
- Public challenge system · no conventional CI
- Limit
- Synthetic/reference platform; scenario metrics are assumptions.
05Agent research
AlphaQuant
Docker stack, multi-service architecture, terminal and UI media
- Status
- Engineering demo · no CI
- Limit
- Research and engineering demo; not a trading product.
06Document intelligence
Local Document AI / Ollama PoC
Docker, local inference path, sample workflow and test scripts
- Status
- Public PoC · no CI
- Limit
- Synthetic/sample-document proof; OCR and model support are bounded.
Explore / 16
More from the lab.
Prototypes, earlier builds, and foundations that informed the later work. Repository status reflects the July 2026 audit.
Applied AI
PrototypeAI Cocktail Assistant
Public app and homepage
Prototype-level evidence.
Inspect ↗Predictive healthcare
Historical notebookAzure LOS Prediction
LOS regression on Microsoft toy data
No CI or detected license.
Inspect ↗Local AI
ExperimentChatGPT Local
Local/containerized implementation
Limited validation evidence.
Inspect ↗Retrieval
Prompt Flow demoChatWithWiki Azure ML
Azure Prompt Flow RAG implementation and README
No tests, CI, or detected license.
Inspect ↗Infrastructure
Learning experimentDocker Python App
Container and Kubernetes implementation
Minimal project depth.
Inspect ↗Document intelligence
CI currently redFreshTrack AI Module
Rich API, tests, Docker, docs, synthetic sample
Default-tip test, security, and build checks are failing.
Inspect ↗Healthcare analytics
PrototypeHEDIS Dashboard
Streamlit surface
README is template-like; no CI or tests.
Inspect ↗Scientific computing
Historical visualizationHNSC Visualization
Authentic slide-image activation visualization
Small research example.
Inspect ↗Predictive healthcare
ExperimentMedication Recommendation
Public implementation and README
No maturity or clinical-validity claim.
Inspect ↗MLOps
Reference studyAzure MLOps Reference
Architecture and documentation
Implementation study, not proprietary production proof.
Inspect ↗Applied AI
ExperimentNeetcode GPT
Public code and README
Limited validation evidence.
Inspect ↗Forecasting
GitHub-archivedNFL Forecasting
Historical backtests and self-evaluated model comparison
Archived, no CI, self-reported metrics, and no detected license file.
Inspect ↗Foundations
HistoricalPython Projects
Earlier Python practice
No modern validation maturity.
Inspect ↗Scientific computing
HistoricalR Projects
Statistical computing and analysis history
No modern CI maturity.
Inspect ↗Document intelligence
Design documentsRAG for HEDIS
Architecture and proposal artifacts
Not a validated executable system.
Inspect ↗Scientific computing
Historical courseworkW5 Biomedical Work
Biomedical/statistical foundation evidence
Historical learning artifact.
Inspect ↗Archive / 17Historical learning, small utilities, and repositories not promoted as current systemsOpen inventory +
Learning history
30 Days of Code
Coding practice history
ArchivedRepository ↗Archive
Check123
Repository history
ArchivedRepository ↗Archive
CRS
Repository history
HistoricalRepository ↗Retrieval
FAISS Nearest-Neighbor Search
Vector-search exploration
HistoricalRepository ↗Archive
FP
Repository record only
Archived · emptyRepository ↗NLP foundations
TF-IDF Weights
Text-weighting implementation
HistoricalRepository ↗Systems
GPU Analysis
Performance/systems exploration
HistoricalRepository ↗Agent experiments
LangChain with Gemini
Framework experiment
HistoricalRepository ↗Analytics
Market Analyzer App
Application experiment
HistoricalRepository ↗MLOps
ML Templates
Pylint workflow and reusable templates
Template collectionRepository ↗Learning history
Neetcode Submissions
Algorithm practice
HistoricalRepository ↗Bioinformatics
Rosalind
Bioinformatics problem-solving history
HistoricalRepository ↗Applications
Streamlit Apps
UI prototyping history
HistoricalRepository ↗NLP
Summarizer App
Summarization application
HistoricalRepository ↗ML foundations
Titanic Transformed Kaggle
Tabular ML practice
HistoricalRepository ↗Local AI
vLLM Experiment
Serving experiment
HistoricalRepository ↗Automation
Webbot Reference
Automation reference implementation
HistoricalRepository ↗