Entrepreneur in Residence
Terrarium (Company Ventures)
Innovating in healthcare.
About
I work on messy, high-stakes problems where systems are fragmented, incentives are misaligned, and the path forward isn't obvious.
I studied Economics and Data Science at Brown. When I developed a rare health issue that went undiagnosed, I mapped symptoms, analyzed longitudinal data, and used AI-assisted reasoning to uncover the root cause after the healthcare system failed to connect the dots. That experience reshaped how I think about decision-making, incentives, and the real cost of siloed information.
Since then, I've designed and shipped production internal AI and analytics platforms, led operations and product delivery at an LLM startup, and built a company-wide knowledge system still used by 200+ staff at Fannie Mae. My work sits at the intersection of applied AI, analytics infrastructure, and product thinking—turning ambiguous questions into durable systems that shape real decisions.
I take ownership end-to-end, move quickly from idea to production, and gravitate toward problems where leverage, judgment, and trust matter.
Outside of work, I climb, surf, cycle, cook, and spend as much time outdoors as possible with my two dogs.
Experience
Experience
Building and shipping in healthcare AI — from production systems at enterprise telehealth to founding the next generation of patient-facing companies.
Applied AI Engineer
K Health
Built and owned production AI systems at K Health — autonomous agents, real-time anomaly detection, and LLM analytics platforms serving enterprise health system partners.
- Built Hawk-GPT, a production Slack bot serving 6 enterprise health system partners — an autonomous analytics agent that diagnoses operational problems through multi-step investigation (causal metric graph, sub-queries, root-cause synthesis). Combines hybrid RAG (semantic + BM25 over 794 schema fields), template matching, self-correcting query execution, and full eval/observability pipeline. Non-technical ops staff now self-serve analytics that previously required analyst support.
- Designed and deployed a real-time anomaly detection platform monitoring the full telehealth patient journey — video calls, intake, SMS, registration, surveys — across 6 partners. 15 statistical checks (z-score anomaly detection against 4-week seasonal baselines), LLM transcript classification with session recovery detection, and a calibrated AI overseer validated via backtesting against 34+ historical scenarios. Runs every 15 minutes; cut mean time-to-detection from hours to minutes.
- Built and deployed an autonomous AI agent for ops teams — a config-driven, two-pass LLM platform (Gemini) that analyzes call transcripts for patient sentiment, intent, and resolution patterns. Applied across 13+ campaigns and 4 partners. First analysis identified that 23.7% of inbound volume had the lowest resolution rate, directly reshaping the product roadmap.
- Engineered 15+ Claude Code skills as multi-agent workflows — git branching, validation, metrics comparison, MR generation, and ticket updates — compressing multi-hour processes into single commands.
I build AI-powered health platforms at ThriveAI Labs: HealthDiaryAI for tracking and AI insights, and ThriveAI for biohackers and athletes. Paused as I transition to a full-time role.
- HealthDiaryAI — personal health tracker to log symptoms, habits, meds, and test results; uses AI to surface patterns and next steps. Reached 200+ active users in the first month via targeted outreach and PLG.
- ThriveAI — performance optimization web/iOS app for biohackers/athletes; connects wearables and health apps (Apple Health, Oura, Strava, etc.) and delivers personalized insights through an AI chatbot.
- Designed and deployed full-stack features using Next.js / React Native (Expo), Tailwind, Supabase (Postgres, Auth, Realtime), Firebase/Analytics, OpenAI API.
- Leveraged AI-assisted development to ship production-ready features in weeks, with CI/CD and continuous deployment.
- Built analytics to surface trends from 1,000+ logs, enabling actionable health insights.
Owned cross-functional delivery of feature rollouts, onboarding design, and GTM campaigns — bridging product strategy, engineering execution, and user operations.
- Owned cross-functional delivery of feature rollouts, onboarding design, and GTM campaigns — bridging product strategy, engineering execution, and user operations.
- Built a SQL/Python analytics pipeline to identify churn drivers, leveraging data to inform retention strategy and feature prioritization.
- Designed AI-driven onboarding flows and community programs that reduced early drop-off by 40%.
- Partnered with clients to gather feedback, rapidly iterate, and deploy fixes using AI-assisted development tools.
Business Analyst Intern
Fannie Mae
Built and launched a 'lessons learned' knowledge system now used by 200+ staff across operations teams, reducing repeated errors by 30%.
- Built and launched a 'lessons learned' knowledge system now used by 200+ staff across operations teams, reducing repeated errors by 30%.
- Designed database architecture, tagging schema, and search functions to make operational knowledge reusable.
- Secured adoption by demonstrating real-time ROI via reduced incident rates in live reporting dashboards.
Nonprofit Consulting Intern
PwC
Designed donor strategy for Big Thought nonprofit, improving outreach ROI through strategic analysis.
- Designed a donor strategy for Big Thought (TX-based nonprofit), identifying high-fit backers through capital flow and mission alignment analysis
- Built a prospecting framework using stakeholder mapping and fundraising benchmarks, improving outreach ROI and reducing pipeline attrition
- Diagnosed performance gaps across fundraising and engagement metrics, delivering a targeted playbook that reshaped their investor outreach approach
Technical Research Intern
CFRA Research
Consolidated and analyzed customer data to identify high-value services for strategic decision-making.
- Consolidated and analyzed customer data to pinpoint high-value services, informing strategic decision-making for product development
Projects
Featured Work
Personal projects and applications I've built to solve real problems and explore new technologies.
Autonomous Cold Outreach Engine — End-to-End AI Sales System
Built a fully autonomous, agent-driven sales system that finds local-business prospects, generates a custom website redesign for each one, deploys it to a live preview URL, and sends a personalized 4-email outreach sequence — entirely hands-free. Handles the full pipeline from lead discovery to closed customer with zero human intervention per prospect.
PythonPlaywrightVercel APIGmail APIPrompt Engineering
Key Contributions
- Designed a multi-stage agent pipeline: prospect discovery → site scraping → AI-generated redesign → automated deploy → personalized email sequence → reply handling.
- Generated 130+ unique, custom-built preview sites for real prospects, each tailored to the business's industry, brand, and existing site.
- Closed 20+ paying customers ($350–$750 per project) sourced entirely from the autonomous pipeline.
- Engineered the system to run hands-off: a single command finds a prospect, builds the site, deploys it, and schedules the email sequence.
GPT-5 Multimodal Evaluation — MOVE Program Fellowship
Selected for a competitive fellowship to evaluate and improve GPT-5's multimodal capabilities (text, image, video) post-training. Applied a structured, product-minded approach to identify failure modes, reliability gaps, and high-impact improvement opportunities.
LLM EvaluationMultimodalAlignment & SafetyApplied Research
Key Contributions
- Designed targeted evaluation scenarios to stress-test real-world use cases.
- Produced structured feedback and curated test cases that directly informed fine-tuning for accuracy, usability, and alignment.
- Collaborated with research teams to ensure findings translated into measurable model improvements.
Kalshi Prediction Market Bot
Fully autonomous prediction-market trading system for Kalshi. Discovers mispriced contracts across economic, weather, and event markets, sizes positions with edge-aware risk controls, and executes hands-free on a daily schedule.
PythonKalshi APIFRED / Macro DataBacktestingGitHub Actions
Key Contributions
- Built a multi-strategy engine (probability-mispricing + macro/FRED signals) with realistic backtesting that accounts for Kalshi's exact fee formula.
- Designed an autonomous research loop using agent teams to scout strategies, validate them against historical vintages, and merge winners.
- Engineered look-ahead-safe data pipelines using ALFRED vintages so backtests reflect what was actually knowable at trade time.
- Runs hands-free on a daily cron via GitHub Actions with full PnL logging and per-strategy attribution.
HookWatch — Webhook Monitoring
Webhook failure detection service that monitors incoming webhooks to catch failures before customers complain. Creates proxy endpoints between webhook senders and your actual endpoints.
PythonFastAPISQLAlchemyAlerting Systems
Key Contributions
- Built proxy endpoint system that forwards webhooks while capturing metrics: volume, success rates, and latency.
- Implemented intelligent alerting for volume drops (>50%), error rate spikes, and activity gaps.
- Created failure logging with full payload storage enabling webhook replay for debugging and recovery.
- Designed dashboard with real-time monitoring and API-key authentication for multi-tenant use.
Blog
Thoughts on systems thinking, AI, healthcare, and the art of solving ambiguous problems.