It’s been great fun working with Gergely Orosz on the “AI Engineering in the real world” post that just went live on Pragmatic Engineer, and we’re proud to be featured alongside other great companies working in the AI space.

Helping Gergely with the content, though, made us realise how far our team has actually come. We’ve been so heads-down building that we haven’t shared nearly enough about the lessons we’ve learned.

That’s why we’ve created “Building an AI incident responder” a microsite where we’re collecting stories from our team about what it’s really like to work with AI.

Over the last year we’ve been building an incident investigator that can understand and remediate issues alongside you. We’re working at the sharp edge of what’s possible, which has pushed us in many ways: we’ve needed to build internal tools to tackle non-deterministic systems, rethink our product process, and develop resilience while working in ambitious R&D cycles.

We’ve shared these lessons for two reasons: first, other teams can benefit from seeing how we’ve tackled these problems, as they generalise across AI products. But selfishly, working on these challenges is really hard, and we’re trying to find people who are excited to help us solve them.

Sharing what we’ve been up to feels like the perfect way to solve for both. Our team will be sharing their individual posts with commentary over the next few weeks - keep an eye out! I promise the content is really something.

Check out the link in thread👇