Speakers
Göran Krampe
Trustly
Opening Keynote
Göran kicks off the second spring edition of Trustly.dev with a look at where we are and where we're heading with AI, availability, and databases.
Jonas Birmé
Eyevinn Open Source Cloud
CTO at Eyevinn Open Source Cloud
From 20 to 200 PRs a Week: How We Shipped Backlog Zero
We build and operate Eyevinn Open Source Cloud, a platform where AI helps you build and deploy solutions that run on top of ready-made open source services. To keep pace with our own backlog, we redesigned how we develop it: AI agents pick up open issues, write code, open PRs, and stop. Humans review and merge. That one constraint took us from 20 to 200 PRs a week.
This is a concrete implementation story from the team behind Open Source Cloud: what we built, what broke, and why the bottleneck turned out to be reviewing and validating what the agents produce. No roadmap slides.
Martin Stidelius
Trustly
Experience Report
OpenCode Today, probably not tomorrow
Tailor your workflow and allow flexibility with open source AI harness.
Martin shares a first-hand report on using OpenCode – an open source AI coding harness – and the lessons learned around tailoring your own agentic workflow. As the tooling landscape evolves at breakneck speed, staying flexible and owning your setup matters more than ever.
Rouwayd Hanna
Trustly – End User Connectivity
Experience Report
From Days to Hours: Building Bank Integrations with AI Agents
We made our codebase AI-ready, then let different AI agents implement a full bank integration from scratch. What normally takes days the AI managed to do under two hours — here's how we did it and what broke along the way.
Andy Koegh & Tanja Cervilla
Datadog
Andy Koegh, Solutions Engineer • Tanja Cervilla, Sr Customer Success Manager
Datadog: Making Reliability Actionable with SLOs, Developer Portals & MCP
Availability targets only mean something when every engineer can see, understand, and act on them in real time. This session covers practical approaches to setting up and tracking SLOs that actually stick, how to use Datadog's Internal Developer Portal to give engineers a single place to surface everything that matters to their services, and how the Datadog MCP server lets developers query and interact with their observability data without ever leaving their development workflow.
Mark Peacock
Yugabyte
Mark is a Technical Director at YugabyteDB, working hands-on with customers adopting YugabyteDB. He is frequently onboarded directly into customer engineering environments to help migrate, tune, and scale PostgreSQL workloads on YugabyteDB, including schema design, SQL optimisation, indexing, and general application performance improvements.
Modernizing and Scaling PostgreSQL with YugabyteDB
Mark explores how YugabyteDB enables modernization and scale for PostgreSQL applications, while preserving PostgreSQL compatibility and avoiding major redesigns. The session covers distributed SQL data modelling, online operations, multi-cloud resilience, and multi-tenancy capabilities.
The talk focuses on balancing high write scalability with efficient query patterns, simplifying PostgreSQL operational challenges such as upgrades and schema changes, and enabling low-latency cross-cloud and cross-region fault-tolerant single database deployments with noisy neighbour protection.
- Distributed SQL data modelling
- Ordered LIMIT scans and hotspot avoidance
- Rolling PostgreSQL major upgrades
- Online DDL
- Low-latency multi-cloud RPO=0 deployments
- Multi-tenancy and noisy neighbour protection
Mikael Wikblom
Trustly – Recurring Experience
Experience Report
Postgres to Yugabyte – what we learned so far
A practical summary of what we discovered while migrating from Postgres to Yugabyte, with key application and database learnings that can help other teams planning the same move.
Bishal Thapa
Trustly – Local Payment Methods & Payout
Experience Report
Testing Resilience with Toxiproxy: Breaking Things Before Production Does
Bishal shares how Toxiproxy was used to simulate failures — database outages, network latency — in a controlled way, revealing how services behave when dependencies are slow, unstable, or unavailable.
He'll also show how they used it both during local development and in automated integration tests, giving a simple and reliable way to validate retries, fallbacks, and recovery flows before those failures happen in production.