Jeff Rogers · Atriumn

Engineering leadership for teams building with AI.

Twenty-five years building software and the teams behind it, at Thoughtworks, Groupon, Angie's List, Centro, Tempus AI and Pinger. Today I work with founders and engineering leaders as a fractional CTO, advisor or hands-on builder, and I build my own products, starting with Idynic.

Book an intro call See what I'm building

426 PRs merged this week by my agent build loop

Live from the build loop

Shipping right now.

Issues go in, reviewed pull requests come out. Every row is a real build by the agent pipeline I run, with what it cost and how many tries it took.

How the loop works →

PRs merged, last 7 days: 426
agent builds shown: 39
verified on the first try: 10%
median compute per build: $0.85

Tue 1:16aRunbookRedesigned a screen layoutredesignsonnet15 tool calls0 test runs≈$0.26revising
Tue 12:36aRunbookRedesigned phone screensredesignsonnet98 tool calls0 test runs≈$2.25revising
Mon 11:25pRunbookRedesigned mobile screensredesignsonnet≈$0.00failed
Mon 11:20pRunbookRedesigned phone screensredesignsonnet24 tool calls0 test runs≈$0.45revising
Tue 12:00aRunbookRefined the design systemdesign-systemsonnet13 tool calls1 test run≈$0.22verified · try 4
Mon 10:56pRunbookRedesigned a screenredesignsonnet24 tool calls0 test runs≈$0.33revising
Mon 10:16pRunbookRedesigned mobile screensredesignsonnet103 tool calls0 test runs≈$2.32revising
Mon 9:40pRunbookRedesigned a main screenredesignopus101 tool calls9 test runs≈$3.90verified · try 4
Mon 9:05pRunbookRedesigned a screen layoutredesignopus25 tool calls1 test run≈$0.90revising
Mon 10:25pRunbookAdjusted design systemdesign-systemsonnet9 tool calls1 test run≈$0.15revising
Mon 6:38pIdynicAdded an integration stepfeatureopus7 tool calls0 test runs≈$0.37revising
Mon 6:12pRunbookShipped a featurefeatureopusfailed
Mon 8:00pRunbookRedesigned a main screenredesignopus20 tool calls1 test run≈$0.60revising
Mon 6:25pRunbookPolished stylingpolishsonnet27 tool calls0 test runs≈$0.34revising
Mon 6:05pRunbookRedesigned a main screenredesignopus253 tool calls22 test runs≈$13.87revising
Mon 5:30pRunbookShipped a featurefeatureopus15 tool calls1 test run≈$0.57verified · try 3
Mon 4:35pRunbookAdded a settings screenfeatureopus22 tool calls2 test runs≈$0.50revising
Mon 6:12pRunbookShipped a featurefeatureopus68 tool calls4 test runs≈$1.90revising
Mon 1:45pRunbookAdded a settings stepfeatureopus105 tool calls3 test runs≈$4.20revising
Mon 6:13pIdynicAdded an output workflowfeatureopus55 tool calls1 test run≈$1.37revising
Mon 4:45pRunbookAdded ai generation for new contentfeatureopus26 tool calls4 test runs≈$0.72verified · try 2
Mon 11:01aIdynicAdded an output stepfeatureopus26 tool calls0 test runs≈$0.85revising
Mon 6:04pRunbookUpdated a design systemdesign-systemsonnet41 tool calls1 test run≈$0.70revising
Mon 3:25pRunbookShipped a featurefeatureopus71 tool calls6 test runs≈$2.35verified
Mon 10:51aRunbookAdded AI output generationfeatureopus161 tool calls9 test runs≈$7.94revising
Mon 10:50aIdynicAdded a new stepfeatureopus94 tool calls7 test runs≈$3.49revising
Mon 6:13pIdynicImproved caching for ai outputupgradeopus82 tool calls8 test runs≈$2.30revising
Mon 6:13pIdynicAdded an integrationfeatureopus16 tool calls0 test runs≈$0.45revising
Mon 7:55aIdynicAdded an authentication flowfeatureopus35 tool calls6 test runs≈$0.85revising
Mon 8:25aRunbookAdded new screensfeatureopus18 tool calls1 test run≈$0.62verified · try 4
Mon 11:05aIdynicFixed a screen performance issuefixopus82 tool calls5 test runs≈$2.64verified · try 4
Mon 7:15aIdynicAdded an integration stepfeatureopus31 tool calls6 test runs≈$0.80revising
Mon 8:01aIdynicExtended caching across ai generationfeatureopus45 tool calls1 test run≈$1.98revising
Mon 7:15aRunbookAdded new screensfeatureopus72 tool calls2 test runs≈$3.14revising
Mon 9:05aIdynicAdded an output stepfeatureopus11 tool calls0 test runs≈$0.44verified · try 2
Mon 7:00aIdynicFixed a screen issuefixopus18 tool calls0 test runs≈$0.73revising
Mon 6:30aIdynicAdded data handlingfeatureopus41 tool calls0 test runs≈$0.86verified · try 3
Mon 5:50aIdynicFixed a page stability issuefixopus42 tool calls1 test run≈$1.20revising
Mon 6:55aIdynicImproved caching for ai outputupgradeopus38 tool calls2 test runs≈$2.19revising
Mon 3:39aBuild loopRan the build loop's self-testself-testhaiku16 tool calls2 test runs≈$0.01verified · try 2

Tue 1:16aRunbookRedesigned a screen layoutredesignsonnet15 tool calls0 test runs≈$0.26revising
Tue 12:36aRunbookRedesigned phone screensredesignsonnet98 tool calls0 test runs≈$2.25revising
Mon 11:25pRunbookRedesigned mobile screensredesignsonnet≈$0.00failed
Mon 11:20pRunbookRedesigned phone screensredesignsonnet24 tool calls0 test runs≈$0.45revising
Tue 12:00aRunbookRefined the design systemdesign-systemsonnet13 tool calls1 test run≈$0.22verified · try 4
Mon 10:56pRunbookRedesigned a screenredesignsonnet24 tool calls0 test runs≈$0.33revising
Mon 10:16pRunbookRedesigned mobile screensredesignsonnet103 tool calls0 test runs≈$2.32revising
Mon 9:40pRunbookRedesigned a main screenredesignopus101 tool calls9 test runs≈$3.90verified · try 4
Mon 9:05pRunbookRedesigned a screen layoutredesignopus25 tool calls1 test run≈$0.90revising
Mon 10:25pRunbookAdjusted design systemdesign-systemsonnet9 tool calls1 test run≈$0.15revising
Mon 6:38pIdynicAdded an integration stepfeatureopus7 tool calls0 test runs≈$0.37revising
Mon 6:12pRunbookShipped a featurefeatureopusfailed
Mon 8:00pRunbookRedesigned a main screenredesignopus20 tool calls1 test run≈$0.60revising
Mon 6:25pRunbookPolished stylingpolishsonnet27 tool calls0 test runs≈$0.34revising
Mon 6:05pRunbookRedesigned a main screenredesignopus253 tool calls22 test runs≈$13.87revising
Mon 5:30pRunbookShipped a featurefeatureopus15 tool calls1 test run≈$0.57verified · try 3
Mon 4:35pRunbookAdded a settings screenfeatureopus22 tool calls2 test runs≈$0.50revising
Mon 6:12pRunbookShipped a featurefeatureopus68 tool calls4 test runs≈$1.90revising
Mon 1:45pRunbookAdded a settings stepfeatureopus105 tool calls3 test runs≈$4.20revising
Mon 6:13pIdynicAdded an output workflowfeatureopus55 tool calls1 test run≈$1.37revising
Mon 4:45pRunbookAdded ai generation for new contentfeatureopus26 tool calls4 test runs≈$0.72verified · try 2
Mon 11:01aIdynicAdded an output stepfeatureopus26 tool calls0 test runs≈$0.85revising
Mon 6:04pRunbookUpdated a design systemdesign-systemsonnet41 tool calls1 test run≈$0.70revising
Mon 3:25pRunbookShipped a featurefeatureopus71 tool calls6 test runs≈$2.35verified
Mon 10:51aRunbookAdded AI output generationfeatureopus161 tool calls9 test runs≈$7.94revising
Mon 10:50aIdynicAdded a new stepfeatureopus94 tool calls7 test runs≈$3.49revising
Mon 6:13pIdynicImproved caching for ai outputupgradeopus82 tool calls8 test runs≈$2.30revising
Mon 6:13pIdynicAdded an integrationfeatureopus16 tool calls0 test runs≈$0.45revising
Mon 7:55aIdynicAdded an authentication flowfeatureopus35 tool calls6 test runs≈$0.85revising
Mon 8:25aRunbookAdded new screensfeatureopus18 tool calls1 test run≈$0.62verified · try 4
Mon 11:05aIdynicFixed a screen performance issuefixopus82 tool calls5 test runs≈$2.64verified · try 4
Mon 7:15aIdynicAdded an integration stepfeatureopus31 tool calls6 test runs≈$0.80revising
Mon 8:01aIdynicExtended caching across ai generationfeatureopus45 tool calls1 test run≈$1.98revising
Mon 7:15aRunbookAdded new screensfeatureopus72 tool calls2 test runs≈$3.14revising
Mon 9:05aIdynicAdded an output stepfeatureopus11 tool calls0 test runs≈$0.44verified · try 2
Mon 7:00aIdynicFixed a screen issuefixopus18 tool calls0 test runs≈$0.73revising
Mon 6:30aIdynicAdded data handlingfeatureopus41 tool calls0 test runs≈$0.86verified · try 3
Mon 5:50aIdynicFixed a page stability issuefixopus42 tool calls1 test run≈$1.20revising
Mon 6:55aIdynicImproved caching for ai outputupgradeopus38 tool calls2 test runs≈$2.19revising
Mon 3:39aBuild loopRan the build loop's self-testself-testhaiku16 tool calls2 test runs≈$0.01verified · try 2

Hover to pause. Failures and send-backs stay in the feed.

How I help

Fractional CTO / VP Engineering

Part-time technical leadership: strategy, hiring, architecture and delivery, two to three days a week.

Best for: Series A-C startups scaling from 5 to 50 engineers

Technical Advisory

Regular sessions with founders and engineering leaders on the decisions that shape the team and the platform.

Best for: Founders and eng leaders navigating critical inflection points

Project Consulting

Time-boxed engagements on one hard problem: an AI integration, a migration, a delivery bottleneck.

Best for: Teams facing a specific technical challenge with a deadline

How engagements work →

Now building

Products and platforms I'm building at the intersection of engineering and entrepreneurship.

Everything I've built →

Idynic

Founder

Active

product

An AI career-identity platform. It turns a work history into an evidence-backed profile, then generates tailored résumés, cover letters and recruiter-facing pages from it. Solo-built, and this site runs on it.

Learn more →Visit site ↗

Sabir Homes

Founder

Pre-launch

service

A home maintenance and concierge service for homeowners.

Learn more →

Ultimate Ninjas

Co-Owner

Active

service

A youth fitness business I co-own, where I run program operations, curriculum and customer experience.

Learn more →

This site is generated live from my Idynic profile.

What Clients Say

A few words from the people I've had the privilege of building with.

AI application development

Jeff has been an invaluable technical partner. The AI-powered tools he built helped us demonstrate real-world value to customers faster than I thought possible. He understands the sales cycle and delivers production-quality software on startup timelines.

Matt P.

Founder & Lead Investor · Sabir Homes

Custom platform development

I came to Jeff with a messy set of manual workflows and walked away with a platform that actually runs my business. He asked the right questions, moved fast, and built something my team adopted immediately. Worth every dollar.

Jamiel S.

Founder · Property Management

Fractional engineering leadership

Jeff joined us as a fractional technical lead at a critical moment — we were under deadline pressure to deliver a government proposal platform and needed someone who could architect it right and get it across the finish line. He did both, and the quality of the work gave us real credibility with our first clients.

Brian R.

CEO · GovTech Startup

AI advisory

Jeff has been an indispensable AI advisor for our team. He cuts through the noise and keeps us oriented on what actually matters in the AI ecosystem — the right tools, the right timing, and how it all applies to our business. His guidance has shaped how we think about our product roadmap.

Justin U.

CEO · Document Storage

Have a team or a product that needs a senior hand?

Book an intro call