Pini Shvartsman
Technology leader. AI transformation and autonomous engineering.
The agent writes the code. The engineer owns the outcome.
I build the systems, the skills and the leadership a company needs to run on AI. I don’t predict that future. I run it in production, and here I share what works, what breaks and what it costs.
82 field notes written since September 2025, from inside a real engineering org.
A 14-part series on rethinking the software lifecycle for agents.
Employee #5 at a startup that grew into the global SaaS platform where I work today.
New here? Start with these three.
The notes that explain how I think. About 25 minutes in total.
- 1
Agentic Overwatch: Why Your Next Dev Team Will Look Like a NASA Control Room
Agentic Overwatch is the discipline of supervising a fleet of autonomous AI agents across every function in production, the way a NOC or SOC supervises live infrastructure: continuous, tiered, shift-based. A reference definition, a maturity model, and a practical starting point for engineering leaders.
- 2
I'm Pro-AI. That's Exactly Why I'm Worried About Our Next Senior Engineers
Practical advice for managers developing junior talent and for junior engineers who want to build real expertise that goes beyond prompting and tool orchestration.
- 3
AI Makes Code Cheap to Produce. Not Cheap to Own.
The governance gap in AI-assisted development: code production is accelerating faster than review, testing, scanning, and ownership can keep up. What the data says and what engineering teams should change.
Prefer the whole argument in one scroll? Take the Briefing.
Rethinking the SDLC
The software development lifecycle was built to ration scarce human execution. Agents inverted that. Fourteen parts on what breaks, and what replaces it.
About 3.5 hours of reading. Best read in order.
Start with part 1- Execution Is No Longer the Constraint15 min
- A Ticket Is a Lossy Compression of Intent13 min
- The Scarce Skill Isn't Writing Code. It's Knowing What Shouldn't Exist.13 min
- Human Review Is Scarce Now. Stop Spending It Line by Line.15 min
- Coverage Was Always Theater. Agents Made It Free.13 min
- Your Pipeline Runs Scripts. It Should Be Making Decisions.13 min
- Your Infrastructure Assumes a Human at the Keyboard15 min
- Stop Asking Whether Agents Should Act in an Incident. Decide How Far.14 min
- Your Dashboards Were Built for Eyes. Your Operator Doesn't Have Any.14 min
- Technical Debt Just Got Cheaper. Something Worse Moved In.14 min
- DORA Measured the Old Constraint. Agents Broke the Instrument.15 min
- Your Agent Has Prod Access. Your IAM Thinks It's a Person.15 min
- You're Not Managing Engineers Anymore. You're Allocating Judgment.13 min
- The Agentic Software Factory: Every Control Surface, Named14 min
Latest notes
Newest first, outside the series above. No prompt tips, no hype.
Everyone Trained Their Engineers on AI. The Gap Didn't Move.
12 minMake the Cheap Path the Default. Make the Expensive Path Prove It.
13 minAlberta Scanned 466 Million Lines of Code in 20 Hours. The Architecture Is the Story.
4 minStop Reviewing Code. Start Reviewing Evidence.
9 minThe One-Man Show Company. Don't Let the Monkeys Touch Production.
14 min
Things I build
Free tools I built to fix problems in my own work.
Open source
MergeStory
Turns a pull request into a short video that explains what changed, why, and what to review.
Open source
aidentity
Keeps an admin, a test and an everyday account of the same Mac app apart, each in its own window.
Who writes this
I have worked in tech since I was 22: IT support, NOC, full-stack, DevOps, architecture and management. I joined a startup as employee #5 and built the CI/CD, the infrastructure and the first offshore team.
That startup is now a global SaaS platform. I have changed how it ships software twice: first with pipelines, now with agents. Today I drive its AI agenda, and what I learn, I share here.
More about me