AI code review on every pull request
I added a GitHub Actions workflow that has Claude review every pull request and post its findings. It informs review rather than replacing it: a person still approves and merges.
GitHub ActionsClaude Code Action
Open to full-stack and mobile engineering roles
Full-stack and mobile engineer. I ship subscription apps.
I’m Jon Zaro, lead developer on Burn by Cara Loren: the iOS and Android app, the API, payments, analytics and the release pipeline. I build with frontier AI models every day, from Claude Code in my editor to agent workflows that review pull requests and automate reporting.
- Based in Seattle, working globally
- Lead developer since Jul 2024
- TypeScript · React Native · Node · Postgres
- Builds with AI every day
Shipped at Burn
Burn by Cara Loren · iOS, Android, web
What changed at Burn
- Releases12 hrs → 5 minTime to recover from a bad releaseA broken update once took 12 hours to fix. Now it rolls back in minutes.Read the case study
- Analytics80% → 47%The team’s key growth number, correctedChecked against billing data, the real conversion rate was far lower than the one being used.Read the case study
- Billing data90% fewerSubscriptions stuck in a failed-payment stateRepaired live billing records in place, with an undo for every change.Read the case study
- Integrations700+Stuck signup and payment tasks clearedReplaced a frozen Zapier queue with tested webhooks in our own API.Read the case study
From the people I’ve worked with
“He picks up complex systems and tools incredibly quickly — and doesn’t just learn them, he masters them and starts making an impact right away. … Our team runs more efficiently and has more time to focus on revenue growth because of his contributions.”
Victoria Lynn C. “He consistently took projects from concept to deployment, seamlessly handling both frontend and backend tasks. His expertise in AI tooling, particularly integrating custom GPT models into apps, made a huge impact on project efficiency and functionality.”
Steve R. “Jon is a strong engineer with a unique ability to think outside the box. While working with Jon, he was crucial in business automations utilizing various tools including AI. These automations led to saved time and efficiency as a company.”
Kevin M.
Experience
Three years shipping production code.
Lead Web Developer · Smash Creative
Jul 2024 – PresentLead engineering on Burn by Cara Loren, a subscription fitness app on iOS, Android and the web. I own the architecture, the API, payments, analytics and the release pipeline.
- Built the release safety net: CI gates, one-click rollback and migrations with tested down scripts.
- Added AI code review to CI and systematized AI-powered customer service processes.
Web Developer · Smash Creative
Jan – Jun 2024Built and released v3.0 of the Burn iOS and Android app, including a progressive difficulty algorithm that adjusts workouts to each user’s performance.
- Built an Instagram tracking automation that cut manual data collection by 80%.
Web Development Intern · Smash Creative
Aug – Dec 2023Refactored the revenue reporting dashboard, cutting report prep from 2 hours to 20 minutes, and built Airtable and Zapier automations for operations.
Volunteer Developer · Lehi Fire Department
Jan 2022 – PresentPart of the team building a protocols app that gives firefighters fast access to procedures, drug dosages and pediatric emergency information.
Full-stack JavaScript certificate · V School
2023
Skills
- Mobile
- React NativeExpoEAS Build & UpdateApp Store & Play releases
- Web
- TypeScriptReactAstroTailwindHTML & CSS
- Backend & data
- NodeExpressPrismaPostgreSQLSQLMongoDBFirebase
- Payments & analytics
- StripeRevenueCatWebhooksPostHogFeature flagsA/B tests
- Infra & release
- GitHub ActionsFly.ioVercelNetlifyDopplerSentry
- AI & automation
- Claude CodeAgent workflowsSubagentsHooksCustom GPTsZapierAirtable
AI and automation
I build with frontier models every day, and I verify what they ship.
AI is part of how I engineer, not a side experiment: in code review, in how I run long tasks, and in the tooling that keeps my work organized between sessions.
An agent operations system for my own work
Claude Code hooks and scheduled jobs brief each session with open work, write handoffs between sessions, turn meeting notes into action items, and generate a weekly digest and pre-meeting prep. Written in Python, with tests.
Claude Code hooksPythoncron
Multi-agent delegation
A lead model plans and reviews while smaller, cheaper models run searches and scoped coding tasks in isolated git worktrees. I read every diff and its test output before it lands.
SubagentsGit worktreesModel routing
Workflow automation
Zapier and Airtable pipelines for operations and sales, an Instagram tracking tool that cut manual collection by 80%, and later moving a frozen Zapier pipeline into tested first-party webhooks.
ZapierAirtableWebhooks
Case studies
Five problems from inside one production app.
Burn is a subscription fitness app on iOS, Android and the web. These are the problems I was trusted with there, from data repair to release safety, and what changed because of them.
Aug – Sep 2026
Repairing live subscription data, with an undo button
A backfill left subscription data double-counting paying users. I contained it, repaired it in production, and made every repair reversible.
14.9% → 6.1%Overstated active subscriptions after the first repair pass
TypeScript · Prisma · PostgreSQLRead case study
May – Jul 2026
Replacing a frozen Zapier pipeline with first-party webhooks
Signup and payment data ran through Zapier, and the queue had been frozen for two months. I moved it into the API with signed webhooks and tests.
700+Stuck Zapier tasks replaced by direct webhooks
Express · Prisma · PostgreSQLRead case study
May – Aug 2026
A launch crash, a postmortem, and a CI gate so it can't happen twice
An over-the-air update crashed the app on launch. I hotfixed it, wrote the postmortem, and built the gate and rollback tooling that stop a repeat.
< 5 minProduction rollback, down from a manual rebuild
Expo EAS Update · React Native · GitHub ActionsRead case study
May – Aug 2026
The activation metric was wrong. I rebuilt analytics on ground truth.
The team was steering by an activation threshold and trial counts that were both overstated. I re-checked them against billing data and rebuilt the metrics.
−33 ptsCorrection to the headline activation metric
PostHog · RevenueCat API · SQLRead case study
Sep – Oct 2026
Moving a live checkout site off WordPress, verified with real purchases
I ran the cutover of the signup and checkout site from WordPress to Vercel, and proved purchase, app access and cancellation worked with real money.
17 / 17Production paths verified on cutover day
Astro · Vercel · Stripe CheckoutRead case study
Hiring for problems like these?I’m open to full-stack and mobile engineering roles.Email me How I work
Four rules I don’t break.
- 01
Measure before I build
Most expensive bugs I’ve fixed were wrong numbers, not crashes. I reconcile against a second source before I trust a metric.
- 02
Every repair is reversible
Production data changes record their own undo. Releases have a rollback that takes minutes, not a rebuild.
- 03
Merged isn’t shipped
A fix counts when the deployed revision includes it and the data shows the problem stopped, not when the PR merges.
- 04
AI-accelerated, human-verified
I build with frontier models every day, and every change still ships with tests that fail without the fix.
Side projects
Things I built to learn.
Developer Dashboard
A focused activity dashboard that pulls actionable work into one view. React, TypeScript, Vite and Tailwind.
PrizePortal
A mock iOS loyalty app with authentication, point tracking and gift redemption. Expo and React Native.
HopeBot
A custom-trained GPT recovery coach with a no-code landing page and a Stripe paywall.
Hiring a full-stack or mobile engineer?
I’m based in Seattle, working globally, and open to full-stack and mobile roles across apps, payments and the systems behind them. Email is the fastest way to reach me.