Developer Productivity @ Spotify

Background

I joined Spotify's Platform Developer Experience area, working on the Spotify Plugins for Backstage and Spotify Portal for Backstage initatives. I helped work on bringing parts of the internal platform that thousands of Spotify engineers use every day to engineers outside of Spotify, as part of Spotify's first big bet on a non-audio, high-margin revenue stream. At the beginning of my time at Spotify, my team's focus was on plugin infrastructure and developer tooling, but as AI started reshaping how engineers worked, my role shifted along with it. I found myself solving problems at the bleeding edge of LLMs and developer experience in production - building AI agents, context marketplaces, evaluating and improving MCP quality across the company at scale, etc.

Context

I'm most proud of helping launch an improved version of AiKA, Spotify's internal AI assistant embedded in Backstage. As LLMs took off in adoption, a RAG version of a general-purpose AI assistant quickly became one of Spotify's hottest and most used tools for users of all disciplines. As the field evolved, MCPs started to proliferate around the company - but with no popular clients to really bind them all together to solve common workflows. Naturally, we turned to AiKA and wondered what a more agentic version of the system would look like. I helped drive that shift while contributing a broad set of product and platform improvements along the way - and the switch to an agentic client unlocked the ability for the hundreds of MCP servers that were written across the company to be integratable with AiKA. With AiKA being the surface where over 80% of the company executed their workflows from, it was the project that most directly shaped how employees interacted with AI at Spotify.

A little while after, I helped build the AI Explorer plugin for Backstage. When agentic AI tooling started to become more widely supported by agents, hundreds of teams across Spotify were creating AI skills, rules, and MCP servers, but there was no way to consistently find them, no way to evaluate their quality, and no standard for how they should work. I helped designed a system that would catalog these different types of AI context into one single pane of glass for every developer.

Final thoughts

Working at Spotify has been a dream. I always wanted to work here since I was a kid, and platform developer experience was the exact area I had wanted to work in. Spotify had always had an exceptionally strong culture of developer empathy, and prior to working at Spotify, I had noticed that I loved creating platforms and tools that made the people around me happier. When the rise in LLMs coming together with my lifelong dreams of being able to leverage AI in fun and new ways, my work at Spotify really does represent a combination of happy coincidences that have let me grow so deeply as a software engineer.