The Story

The Mission

IT career background. No engineering degree. Just systems thinking, a lot of reading, and Claude as a full-time engineering partner.

I've spent twenty years making complex systems reliable. In February 2026, I turned that thinking on myself — and built the platform this site runs on.

My job has always been making complex systems reliable and getting people to actually use them. I'm not a developer by training — I'd never written and deployed production application code, though I've shipped plenty of infrastructure changes, database updates, and network upgrades. But I understand systems, and I understand what it takes to change behavior.

In February 2026, I decided to apply that same thinking to myself. Not a challenge. Not a hack. A proper system — with data sources, intelligence layers, automated coaching, and public accountability.

I built the Life Platform as a solo project: 26 data sources, 121 intelligence tools, 62 Lambda functions, and an AI that writes me a coaching brief every morning. The whole thing runs on AWS for ~$19/month. Every number is real. Every failure is included.

This project also became a proof of concept for something I care about: what happens when someone with a vision and AI development tools builds without limits? The answer, so far, is that they build things nobody expected — including themselves.

Mission brief
What Does Done Look Like?
Physical
Weight 307 → 185 lbs
Lost so far — lbs
Movement Walk, lift, run. Daily.
Mind & Connection
Social Close friends, closer family
Mental health Less noise, more presence
Journaling Consistent — the data needs it
Satisfaction Content. Not performing it.
The System
Character score 7 pillars → 80+/100
Consistency No more 5-month gaps
Status Active
Based in Seattle, WA
Why not just use Apple Health?

Apple Health aggregates data but doesn't analyze it. There's no AI coaching layer, no cross-source correlation engine, no character sheet, no public accountability system, and no weekly narrative journalism. The value of this platform isn't the data collection — it's the intelligence layer on top: the part that notices your HRV dropped the same week you stopped journaling, and asks you why.

Explore the project

What I've built

Origin The Story Where this started, what the first 90 days actually looked like, and why I decided to do it in public. Read → Architecture The Platform 62 Lambda functions, 26 data sources, 121 intelligence tools. How the system actually works. See the stack → Weekly The Chronicle An AI journalist named Elena Voss reads my data every week and writes about what she finds. Read the journal →
Technical context

The build

Role Solo developer. Claude handles all code generation; I handle architecture, product decisions, and every deploy.
Infrastructure AWS — Lambda, DynamoDB, S3, CloudFront, EventBridge, SES, Secrets Manager. Single AWS account, us-west-2.
Intelligence Anthropic Claude via MCP protocol — 121 tools exposed to Claude for live data access. Opus 4 for clinical reasoning, Sonnet 4 for structured outputs, Haiku 4 for lightweight daily compute.
Data sources 26 active — Whoop, Eight Sleep, Withings, Garmin, Strava, Apple Health, MacroFactor, Habitify, Notion, Todoist, Dexcom Stelo, genome, labs, DEXA, weather, travel, blood pressure, supplements, state of mind.
Test coverage 83 tests, all passing. Architecture Review #19: grade A. CI/CD pipeline with auto-rollback.
Cost ~$19/month on AWS. Anthropic API calls are the variable cost; Lambda compute is negligible.
Purpose Personal health intelligence, accountability, and a live demonstration of what one person can build with AI. Started February 2026.
Get in touch

If You Want to Connect

This started as a personal project, and I haven't deliberately built it for an audience beyond the people already in my life. But if something here resonates — if you're going through something similar, if you're curious about the platform, or if you just want to say hello — I'd genuinely love to hear from you.

Say hello matt@averagejoematt.com