

Building with AI at work | Mentimeter Stockholm Meetup
AI agents can write code, but how do you trust it, keep it secure, and know it works? Join us at Mentimeter for an evening of practical talks on building with AI at work, from agent guardrails and software verification to growth experimentation. Come for fresh ideas, useful lessons, and conversation with others working through the same questions.
Doors open at 17:00, with talks beginning at 17:30. We'll take a break for snacks and drinks at 18:00, then continue with pizza and mingling from 19:00. Pizza, snacks, and drinks provided.
Talks
Beyond the coding agent: building a software factory we can trust - Rafael Belliard
AI agents can write code fast. What does it take to trust that code enough that engineers can focus on building the right product, across a whole engineering org? Rafael walks through the lifecycle Mentimeter is building around its coding agents: the choices we leave to people, the work we hand to agents, and how every run makes the next one better.
Security guardrails around an agentic workflow - Loraine Garutti
An agent can make changes long before anyone merges a PR, so each moment has its own guardrails. Inside our Agentic Development Lifecycle, the agent's commands and code have their own guardrails before even writing a PR. After opening a PR, an Infrastructure-as-Code repository handles access management and decides who can deploy to production.
Verification engineering: how we know software works - Jaunius Eitmantis
Code review and a test suite are the usual ways to know a change works. Both get harder to rely on as the amount of code grows, and nobody wants to wait long to find out. This talk looks at what else a team can build, from techniques such as TLA+ to signed-off CI runs on your own machine.
AI across the growth experimentation loop - Oleksii Hopchenko
Mentimeter's Growth Lab learns what helps people start, return, and get value from Mentimeter by running experiments. This talk walks through that loop: a workshop idea becomes an experiment doc, the doc becomes an A/B test, and the result goes back to the company as something the next team can reuse. You'll see where AI drafts the doc, sets up the test, and writes the share-out, and where a person still has to decide.
Event schedule
17:00 — Doors open
17:25 — Welcome
17:30 — Rafael
17:45 — Loraine
18:00 — Break. Snacks and drinks
18:20 — Jaunius
18:40 — Oleksii
19:00 — Pizza and mingling
20:00 — Close
Hope to see you there!
Your host,
Johanna Blom, Product Engineer