I'm an AI-native product partner and squad-of-one. I work with your team so what we build actually ships and sticks.
For B2B / knowledge-work companies who want to ship AI features fast without hiring a full AI team.
"Hourly billing penalizes speed. I charge for shipped features. If I finish in 2 days what takes others 2 weeks, you get it sooner."
Hourly billing incentivizes slowness. I scope work as feature units (points) with fixed pricing.
System Error:
"Billing: 40 hours logged."
Fix: Explicit features shipped every sprint.
Optimization:
I leverage LLMs, RAG, and Agents to multiply output. I bring 20+ years of experience + AI leverage.
Raw AI generates bugs; raw human effort is too slow. I operate at the intersection: AI for generation, Human for architecture and verification.
Technology fails when humans resist it. My background in facilitation means I engineer for adoption, not just latency.
Human Check:
Does the team understand *why* we are building this? Yes.
I use a simple point system: each feature gets a size (XS/S/M/L). You buy a block of points per month, not a block of my time. Clear scope, clear price, no clock-watching.
"Map and ship your first AI win."
A 2–4 week engagement to find and ship a first meaningful AI use case. We map opportunities, pick one, and deliver a working feature.
"Ongoing AI feature velocity."
A dedicated AI engineer-of-one on a points-based subscription. You buy a block of feature points; I pull from your backlog and ship.
"Safety & Strategy."
For teams who have devs but lack AI seniority. I audit, guide, and ensure you aren't building vaporware or security holes.
"Don't just buy the fish."
Workshops to upgrade your team's wetware. From "Zero to RAG" engineering bootcamps to "Prompt Engineering that actually works."
"The 'Break Glass' option."
You hate points? You love hourly billing? Fine. I offer standard hourly consulting, but it comes at a significant premium: my standard rate is 200€/hour.
(it changes for projects I love, that have good impact, that are open source, etc.)
Selected projects from 20+ years of engineering, facilitation, and weird internet experiments.
Lead AI Engineer for Epikast. Built a RAG-based multi-agent system for Medical Affairs.
Co-founder/CTO. Designed privacy-first AI tools for Internal Family Systems (IFS) therapy.
"Safe AI experimentation without exposing sensitive user data."
Language Technologies Engineer at Apple (Cupertino).
If you dig deep enough, you'll find I contain multitudes.
"Stupid NLP tricks that should not work but do." (It was 2020, back in NLP prehistory)
(sadly we did not record it.)Typical first step is a 15 min call to map AI opportunities and see if the sprint model fits your backlog.
Alessandro Pedori © 2026 | Wetwarecraft & Engineering
Italy. Remote-first.