AI Tinkerers Lisbon Demo Night
⚡ Agent Memory, Context Engineering, and Raw Code
Models are good enough. What breaks agents is memory. Every session starts from zero. The agent forgets what it learned yesterday, never heard what the team decided on the call, and doesn’t know a teammate’s agent is already editing the same file.
What to remember, what to forget, and how to pull back the right thing at the right moment: that’s the hardest part of context engineering, and it’s where a lot of builders are working right now.
On the Sunday night before Web Summit opens, AI Tinkerers Lisbon brings together 70 people who build with AI to show, live, how they’re solving it. No slide decks, no product pitches. Working code only.
📅 Event details
Date: Sunday, November 8, 2026, the night before Web Summit (November 9–12)
Time: 6:00 PM – 10:00 PM WET
Venue: HOOD Lisbon, Marvila, 10 minutes from Web Summit (Rua 3 da Matinha, Edíficio Altejo, Armazém 101, 1950-326 Lisboa)
Capacity: 70 builders (curated)
🧠 What we’ll dig into
Memory across sessions: agents that pick up where they left off, across sessions, machines and teammates.
Forgetting on purpose: what’s worth keeping, what goes stale, and how an agent decides.
Recall that works: pulling back the right thing at the right moment, not just the closest embedding.
Memory from conversations: meetings, calls, Slack threads and tickets turned into something an agent can act on.
Skills and playbooks: what agents learn once and reuse.
Memory gone wrong: an agent that confidently remembered the wrong thing, and how it got fixed.
🎙️ Call for demos: show your code
AI Tinkerers has a strict no-slides policy. We want raw, working, even buggy code and live deployments. Demos are 5–7 minutes and focus on the technical side: “How did you build this?”, not “Why should I buy this?”.
If you’ve built an agent that remembers, forgets or recalls in an interesting way, we want to see it. Propose a demo here: https://lisbon.aitinkerers.org/meetup/mu_xd4Z5-Mf8Ag/speaking
⏱️ Format & schedule
6:00 PM: Doors open, drinks
6:30 PM: Main stage demos (5–7 minute live runs)
7:30 PM: Pizza, beer and networking: who’s hiring, who’s looking for a co-founder, who wants an early user
10:00 PM: Doors close
🥽 Who should attend?
Entry is selective by design, to keep the room high-signal. Space is limited to 70, and we ask for a public LinkedIn, GitHub or X profile showing hands-on work with AI.
This event is curated for:
AI Engineers building agents, memory, retrieval and RAG systems in production.
Technical Founders whose product lives or dies by what their agents remember.
Builders in town for Web Summit who’d rather start the week with people who ship than with a keynote.
🤝 Sponsors
Sponsor: SageOx. SageOx is the hivemind for human-agent teams. It records a team’s conversations and decisions once, recalls them in every Claude Code, Codex or Cursor session, and warns you when a teammate’s agent is already editing the same files. https://sageox.ai
Host: Vexa. Vexa is an open-source API for meeting bots. It sends a bot to a Google Meet, Teams or Zoom call and streams the transcript in real time, self-hosted or in the cloud, so agents know what was actually said. https://vexa.ai
Venue: HOOD Lisbon, Marvila.
Organised by AI Tinkerers Lisbon.