Five profiles you shouldn’t miss: engineers, researchers, and designers who build — every day — the systems and experiences that millions of people use without thinking about it.
Every year, Codemotion’s call for papers draws hundreds of proposals. Among the talks accepted for the 2026 Milan edition, a group of speakers stands out — not for the number of logos on their CV, but for their trajectory: paths that cross disciplines, companies, and technology cycles in non-linear ways, producing a kind of perspective that’s hard to find elsewhere.
Here are the five you can’t miss.
1. Jigyasa Grover — ML Engineer @ Uber Ex Twitter/X · Meta/Facebook · Faire · Bordo AI
Jigyasa has built large-scale machine learning systems at some of Silicon Valley’s most demanding companies — moving from Meta to Twitter to her current role at Uber working on AI personalization. She’s the author of Sculpting Data for ML and a member of the Google Developer Advisory Board.
What makes her interesting isn’t the list of names — it’s that she’s lived all of them from inside production systems, not from the outside looking in.
Talk: When NOT to use an agent: Choosing Between Workflows, Services, and LLM Orchestration
Why you shouldn’t miss it: When everyone is pushing toward LLM agents, someone who has built ML at that scale in production already knows where things break. Jigyasa brings a practical framework for deciding whether an agentic architecture is actually justified — weighing latency, costs, observability, and blast radius. You’ll leave with a decision checklist, not more hype.
2. Mario Zechner — Partner @ Earendil · Creator of libGDX
Mario was working in machine learning in the 2000s, before it was cool. Then he built libGDX — the open source Java game framework used by millions of developers — wrote books, moved to San Francisco, came back, sold a startup, ran a game dev business for ten years, and is now building pi, a coding agent. The irony of that arc is intentional.
Talk: Everything is slop, let’s slow the f**k down
Why you shouldn’t miss it: Mario gives an unfiltered account of how coding agents are flooding open source projects with automatically generated issues and PRs. This isn’t a nostalgic talk — it comes from someone who’s building one of those agents. It’s the most honest perspective on the topic you’ll find in the programme.
3. Carolina Pinto — Senior UX Designer @ King (Candy Crush Saga) Ex Dell Technologies · CaixaBank
Carolina isn’t an engineer — and that’s exactly why she deserves your attention. She works on behavioural systems for hundreds of millions of users, designing the mechanics that decide when a game rewards you, when it challenges you, and why you keep coming back. She’s applied that same logic in enterprise contexts at Dell and in financial services at CaixaBank.
Talk: What Candy Crush Can Teach AI Products About Human Behavior
Why you shouldn’t miss it: AI products are becoming increasingly capable but often fail at the experience level. Carolina brings game design principles — progression, emotional feedback, behavioural pacing — and applies them to AI products. It’s the most cross-disciplinary talk in the programme: valuable both for those building developer tools and for anyone thinking about product experience.
4. Giorgio Natili — Head of AI Engineering @ Oracle Cloud Support Ex Amazon · Mozilla · Capital One · OPAQUE Systems
Giorgio’s path is non-linear by design. From Amazon to Mozilla’s open source infrastructure, from Capital One’s financial AI to OPAQUE Systems’ confidential AI — where he built systems enabling inference on sensitive data without exposing it — and now leading AI engineering for enterprise support at Oracle. Each step has brought him closer to the same problem: how to make AI systems reliable where they can’t afford to fail.
Talk: The Ethics Layer: Architecting Sovereign AI and Digital Dignity in Agentic Systems
Why you shouldn’t miss it: This isn’t a talk about AI ethics philosophy. Giorgio proposes a concrete architecture — the Immutable Ethics Policy Layer — as a deterministic control point between agent orchestration and tool execution. If you work on agentic systems in production, this is the most operational talk on the subject.
5. Matteo Collina — Co-Founder & CTO @ Platformatic · Node.js TSC Member Lead Maintainer of Fastify · Creator of Pino
Matteo is the name behind Fastify and Pino, two of the most downloaded JavaScript modules in the world. A Node.js Technical Steering Committee member with a PhD in IoT and over 60 international conference talks to his name, his npm packages are downloaded billions of times a year — by systems you’re probably using right now without knowing it.
Talk: Kill the Standup: How One Engineer Ships What Five Cannot
Why you shouldn’t miss it: An autobiographical and deliberately provocative talk. In three months he wrote 19,000 lines solo for the Node.js Virtual File System — and in doing so introduced a CVE through an AI-assisted PR he didn’t re-read carefully enough. He’ll talk about Conway’s Law, the hidden cost of Scrum, and when working alone is actually the wrong call. It’s rare for someone to bring their own mistakes onto the stage.
🎟️ Want to meet these speakers in person? A unique opportunity to attend their talks, ask questions live, and connect with the Italian and international tech community.
📅 Codemotion Milan 2026 📍 SUPERSTUDIO PIÙ
🎤 18+ speakers, technical talks, workshops and networking
