- π MBA @ UVA Darden '27 - applied AI builder
- π― I build agentic AI products for high-stakes decisions β shipped with the trust layer built in: evidence citations, honest limits, and evals that measure whether the AI is actually right.
- π οΈ Bio/longevity, blockchain, VC and PE, AI-native services, e-learning, etc.
- π« Badwalb27@darden.virginia.edu
![]() |
![]() |
![]() |
| TraceHound | AI Frontier Dispatch | Deal Docket |
| Agentic crypto hack tracer β live Β· executable evals Β· case study | Personalized AI + markets briefing, one-line plugin install β case study on generalizing a personal tool | Deal screening with rank-stability analytics β live Β· case study |
LLM products here ship with tests, not vibes: golden set Β· deterministic policy gate Β· LLM-judge rubric Β· calibration plan (TPR/TNR, bias correction) β each states plainly what's validated vs. designed.
- TraceHound evals β executable deterministic gate (
npm run eval) over a golden set: catches invented hops, false watchlist claims, missing evidence citations, unsafe letter claims. - Consulting Trainer evals β 12-scenario golden set with deliberate tempting-wrong-answers, two-axis LLM judge with anti-halo instructions, calibration plan.
LiveForever (live, private) β personal health control tower unifying wearables (Oura Ring, Apple Health, Whoop), 10+ years of clinical bloodwork & daily self-logged training/habits, and personal genomics (DNA via 23andMe) into one owned data layer.
OpenAI Build Week 2026: Public version - Live β Β· Repo
Benchmarks PhenoAge biological-age, N-of-1 correlation analysis quantifying what each input does to HRV, recovery, resilience scores + 28-day analysis on interventions.
PhaseSignal (Live β) β scores live ClinicalTrials.gov data against a base rate, reweighted across four factors.
TraceHound (Live β) β agentic crypto hack tracer: live hop-by-hop tracing and narration from compromised wallet via Etherscan API, cross-references OFAC watchlist populated with sanctioned addresses from U.S. Treasury's SDN list. Built from experience with federal law enforcement on crypto crime.
Tranche AI (Live β) β condition-gated capital release for VC deals, designed for AI-agent milestone review; smart contract + dispute flow live on Base Sepolia (EAS attestations). Solidity.
Deal Docket (Live β) β deal-screening dashboard built around an AI-enabled service-roll-up thesis; adjustable five-box scoring framework.
AI Stack (Live β) β interactive map of the AI industry from silicon to application layer, value accrual, token cost calculator.
AI Frontier Dispatch β personalizable AI + markets briefing runs inside your harness 3x/week: track frontier builders (Grok API for live X read), GitHub/Hugging Face, deals/careers signals. One-line plugin install.
Orbit (private) β relationship intelligence engine scores tie strength from relationship history, auto-scheduled follow-ups. Obsidian graph + dashboard.
IB Technicals Fluency Trainer β merger-model cockpit, purchase-price allocator, DCF sensitivity heatmap - live playgrounds.
Consulting Case Prep Trainer β profit-diagnosis game, market-sizing builder, exhibit reader. Includes a designed eval harness: golden set, two-axis LLM judge, calibration plan.
The Operator's P&L Room β eight-quarter run-the-business simulator under leverage & covenants, 13-week cash-crisis room for distress-operator decision making.
Short product write-ups β problem Β· users Β· product decisions & tradeoffs Β· how I'd measure success Β· roadmap:
| Project | The product-thinking angle |
|---|---|
| TraceHound β | Agentic AI for underserved users; rigor about a tool's limits |
| AI Frontier Dispatch β | Generalizing a personal tool for others' setups; borrow vs. build |
| Tranche AI β | Scoping a frontier problem in VC |
| OpenAI Build Week '26: LiveForever β | Separating deterministic evidence from model interpretation |
| The AI Stack β | Mapping where value accrues across AI |
| Recruiting Trainer Suite β | Framing three tools as one product line; retention-first design |
| PhaseSignal β | Transparency vs. black-box incumbents |
| Deal Docket β | Making assumptions visible; build-vs-backend judgment |
Most are self-contained apps β vanilla HTML/CSS/JS, no framework, no build step, no dependencies β designed, built, and shipped solo end-to-end. Some split data from rendering (data.json + app.js) for content updates. Some use a real Python data pipeline (data/build_dataset.py) that pulls and scores live data. TraceHound is a Next.js app with server-side API keys.



