The opportunity
The Personalization team makes deciding what to play next easier and more enjoyable for every listener. From Blend to Discover Weekly, we're behind some of Spotify's most-loved features.
What you'll do
Own engineering execution and technical direction for the Conversation: Product Area, including the Agent Engine, the evaluation infrastructure, the Search platform, the AI Voice platform (Speak), and some consumer agentic features built on top of the Agent Engine. This spans production systems at massive scale, applied ML research, and infrastructure that other product areas depend on.
Co-lead the Product Area in a paired product/engineering model. You own: engineering and deeply partner with a product counterpart that owns product strategy. You’ll operate as a single leadership unit for planning, prioritization, and external representation. This partnership is core to the success of the role.
You’ll manage a team of ~70 engineers, with a small team of Senior: Engineering Managers and Staff/Principal engineers directly reporting to you. You’ll develop and hold accountable the EM layer running the PA's roughly 9 teams. Your leverage comes through your SEMs and Staff engineers, but you’re ultimately accountable for delivery across the entire product area.
Drive the migration of the company’s existing agentic infrastructure onto the: shared Agent Engine, in close partnership product teams across the company, balancing migration speed against product teams’ existing roadmap work.
Own the technical strategy for Spotify's first-party voice investment: what Speak builds versus buys, how the team organizes and delivers against a market that's commoditizing fast, and how the character/brand voice decisions for agent experiences get made and shipped on their own aggressive timelines.
Define and hold the technical bar on evaluation. Build the discipline around: building feature-specific judges and datasets and product decisions based on those eval sets.
What they're looking for
- You are deeply technical, not tech-adjacent. This PA's defining work (a new: shared agent runtime and harness, a scaled evaluation system, a voice/speech stack) requires a Director who can evaluate architecture proposals, hold their own in a room with senior ICs, contribute technically in meaningful ways, and make tough tiebreaking and tradeoff calls.
- You have run large engineering orgs through platform consolidations,: migrations, or evolutions, and you’ve had to earn trust by building highly scaled and functional platform technology.
- You have experience with at least two of: LLM-based agent systems and orchestration, evaluation/eval-infrastructure design (including automated judges and regression suites), large-scale retrieval or search systems, or speech/voice ML systems and production voice infrastructure.
- You've managed managers, not just individual contributors. You know how to: build a management structure that scales.
- You have a track record of building strong but nimble teams, with experience: restructuring, shifting team resourcing, adjusting to fast changing priorities or shifting needs without losing delivery momentum.
