Simone Bova / AI engineer, founder, production
The model is a commodity. The method isn't.
I build AI systems that hold up in production. Founder & CTO of Yempik (si apre in una nuova scheda), AI engineer at Intarget (si apre in una nuova scheda), maintainer of code-os. I sit in the room where decisions get made and in the repo where things ship.
This is my public archive: systems built, method, what I think about the AI frontier. No slides. Just work you can inspect up close.
Systems, products, leverage
Work
Yempik
Founder & CTOThe Italian AI software house. The AI others leave you in slides, we put in production: agents, automations, custom software for SMBs and enterprise.
(opens in a new tab)code-os
Maintainer / open sourceThe engineer's operating system for Claude Code: agentic harness, second brain, skills, tech references. The method I work with, shipped as a product.
(opens in a new tab)cowork-os
Co-maintainer / open sourceThe open-source company brain for the business side: decisions, context and operating routines as governable Markdown files. CONTEXT IS GOLD, in practice.
(opens in a new tab)Intarget
AI EngineerThe transformation into a fullstack AI company: a proprietary agentic platform evolving into a governed company brain. Governance and auditability, adoption in the field, training, working sessions with domain experts. From strategic design to implementation.
Freelance delivery
Technical Partner / AIMVPs that reach production: WhatsApp and voice agents with security in the code (not in the prompt), EU data residency, on-call when it matters. A partner, not a ticket-taker.
Operating doctrine, not theory
Method
Model ≠ Problem
If the agent fails, the flaw is in the design. Models have been more than good enough for a while.
"Ignorant" AI-native redesign
Redesign the process from zero under one constraint: one person + these tools. An org chart is not a workflow.
The expert's signature, not the token lottery
Predictable quality via tests on real scenarios and pass^k. The expert's standard lives in the system.
The harness is the asset, the model isn't
Tools, contracts, evals, observability: yours. The model is a swappable component.
Public archive, not a blog
Writing
The smart model is the bottleneck. Treat it that way.
For weeks I've been running a Frankenstein of Claude Code, codex CLI and grok-cli. The smartest model specs and places the seal; everything else is done by something dumber and faster. It's ~30 lines of markdown and it fits in your head.
deploy to prod. make no mistakes.
We launched the first company brain live from the Italian Founders House, surrounded by founders doing their own things. These are the release notes: more sushi than marketing.
The model is a commodity. The method isn't.
95% of GenAI projects produce no ROI. The bottleneck has moved: from the model to the people, the workflows, the infrastructure you own.
Altitude, snow, safety margin
Fuoripista
The mountain punishes the same mistake production does: trusting yesterday's conditions.
Winters on snow, summers on rock. Off-piste skiing and high-altitude hikes are the part of my life where risk assessment isn't a slide: bulletin, checks, margin. Then you drop in, and the fun is real.
They stay here as side notes: present enough to be true, never big enough to become a costume. There's also a story with a helicopter, one night, and a thinner margin than I like to admit. Another time.
Contact