Skip to content
Adi's Digital Garden

Projects

Eval Loops

Active

Coding-agent skills that wrap the LLM-eval pipeline in two hierarchical auto-research loops — so the slow, mechanical parts of evals run themselves while humans are spent only where judgment is irreducible.

LLM Evals Agent Orchestration Claude Skills
2
Orchestrators
4
Self-Improving Loops
6
Human Checkpoints

PHP — Personalized Health Provider

Completed

A trust system for chronic-condition patients between doctor visits — integrating six health products via a knowledge-graph RAG backend, tiered action capabilities, and a doctor-panel eval loop.

Healthcare AI RAG Evals LangGraph
5
Chronic Conditions
Monthly
Eval Cadence
15
Days to Repurpose

Genome360

Completed

Clinical NLP pipeline that mined 4 million oncology patient records to extract genomic biomarker results and surface trial-ready cohorts for pharma partners.

Clinical NLP Oncology Healthcare AI
30K
Patients Surfaced
4M
Records Mined
16.9M→2.5K
Hours Saved

Provenance

Active

Irrefutable proof that writing was created by a human through authentic human processes. Every keystroke, pause, and revision is recorded and cryptographically chained — the proof is the journey, not the output.

AI Writing Tools Cryptography

HIPAA Breach Analysis

Completed

Exploring 15 years of healthcare data breaches reported to the HHS Office for Civil Rights. Interactive visualizations reveal patterns in attack vectors, geographic distribution, and supply chain risk.

Data Visualization Healthcare Security
7,400+
Total Breaches
16
Years of Data
938M+
Individuals Affected