Sign up to save your podcasts
Or
Share Java at Spotify: Microservices, MCP & AI Overload – Mohamed Aboullaite | The Marco Show
Share to email
Share to Facebook
Share to X
Share to email
Share to Facebook
Share to X
Or
Java at Spotify: Microservices, MCP & AI Overload – Mohamed Aboullaite | The Marco Show
Mohamed Aboullaite, backend engineer at Spotify, Java Champion, Google Cloud Developer Expert, and Docker Captain, joins Marco to talk about building AI-powered integrations at scale, what software engineering looks like when you're running five AI agents in parallel, and why the foundations still matter in an AI-first world.
They discuss the engineering behind Spotify's ChatGPT integration (built on MCP), the non-determinism challenges of tool-calling, agentic coding workflows, review fatigue, and a frank conversation about the junior developer pipeline and what it takes to become senior today.
💡 In this episode:
How Spotify's ChatGPT integration works (MCP apps, the Spotify widget inside ChatGPT)
Siri/Alexa/Google Home backends and Spotify's ubiquity strategy
Non-determinism in MCP tool-calling and how Spotify works around it
Running 5 AI agents in parallel: the plan mode, review loops, cognitive fatigue
Java at Spotify: monorepo, microservices, Backstage
AI's impact on junior hiring and how to become senior anyway
Finding mentors and the power of the Java community
The token economy: measuring productivity by tokens burned
⏱️Timestamps:
(00:00) Teaser
(00:50) Meet Mohamed: Spotify backend engineer and Java Champion
(01:53) What Mohamed works on at Spotify
(03:12) Spotify inside ChatGPT and MCP apps
(06:23) Building for new AI platforms
(08:53) Spotify tools, playback, and device switching
(09:48) Tool calling challenges with AI models
(11:24) Using AI in day-to-day development
(13:29) Running multiple coding agents in parallel
(14:39) Why planning matters more than prompting
(16:56) Review fatigue and cognitive load
(19:43) Spotify’s backend, microservices, and Backstage
(21:34) Java’s evolution and the AI era
(24:14) Scala, Kotlin, Haskell, and JVM languages
(25:46) Advice for junior developers in the AI age
(29:27) How to become senior when AI solves everything
(34:09) Finding mentors and growing through community
(37:41) Giveaway question
(39:28) Rapid-fire questions: Morocco, Sweden, Spotify, AI
New episodes every other Wednesday. Subscribe for more developer-focused conversations.
🎥 Watch the full episode on YouTube: https://youtu.be/6WvoouJ9Mrk
View all episodes
By IntelliJ IDEAMohamed Aboullaite, backend engineer at Spotify, Java Champion, Google Cloud Developer Expert, and Docker Captain, joins Marco to talk about building AI-powered integrations at scale, what software engineering looks like when you're running five AI agents in parallel, and why the foundations still matter in an AI-first world.
They discuss the engineering behind Spotify's ChatGPT integration (built on MCP), the non-determinism challenges of tool-calling, agentic coding workflows, review fatigue, and a frank conversation about the junior developer pipeline and what it takes to become senior today.
💡 In this episode:
How Spotify's ChatGPT integration works (MCP apps, the Spotify widget inside ChatGPT)
Siri/Alexa/Google Home backends and Spotify's ubiquity strategy
Non-determinism in MCP tool-calling and how Spotify works around it
Running 5 AI agents in parallel: the plan mode, review loops, cognitive fatigue
Java at Spotify: monorepo, microservices, Backstage
AI's impact on junior hiring and how to become senior anyway
Finding mentors and the power of the Java community
The token economy: measuring productivity by tokens burned
⏱️Timestamps:
(00:00) Teaser
(00:50) Meet Mohamed: Spotify backend engineer and Java Champion
(01:53) What Mohamed works on at Spotify
(03:12) Spotify inside ChatGPT and MCP apps
(06:23) Building for new AI platforms
(08:53) Spotify tools, playback, and device switching
(09:48) Tool calling challenges with AI models
(11:24) Using AI in day-to-day development
(13:29) Running multiple coding agents in parallel
(14:39) Why planning matters more than prompting
(16:56) Review fatigue and cognitive load
(19:43) Spotify’s backend, microservices, and Backstage
(21:34) Java’s evolution and the AI era
(24:14) Scala, Kotlin, Haskell, and JVM languages
(25:46) Advice for junior developers in the AI age
(29:27) How to become senior when AI solves everything
(34:09) Finding mentors and growing through community
(37:41) Giveaway question
(39:28) Rapid-fire questions: Morocco, Sweden, Spotify, AI
New episodes every other Wednesday. Subscribe for more developer-focused conversations.
🎥 Watch the full episode on YouTube: https://youtu.be/6WvoouJ9Mrk