Deepgram is a developer-first speech AI platform offering APIs for speech-to-text, text-to-speech, and voice agents. Its products include Nova-3 for transcription, Aura for text-to-speech, and Voice Agent API for real-time conversational experiences. Deepgram serves both self-serve developers and enterprise customers in contact center, healthcare, and conversational AI. You are an Engineering Manager responsible for one of the platform teams supporting core API capabilities.
Deepgram's leadership is deciding how to align engineering investment for the next 12 months. The company has three competing priorities: (1) accelerate enterprise growth in contact center and regulated industries, (2) improve developer adoption of Deepgram's self-serve APIs, and (3) defend against increasingly aggressive competitors bundling speech AI into broader AI platforms. Engineering capacity is constrained, and recent roadmap discussions have become fragmented: product wants faster feature delivery for Voice Agent API, sales wants enterprise security and compliance work, and developer relations wants lower onboarding friction for Nova-3 and Aura.
As the Engineering Manager, you need to propose a strategy for ensuring engineering work stays tightly aligned with broader product and company strategy rather than becoming a collection of team-level priorities.
| Metric | Current State |
|---|---|
| Annual recurring revenue (ARR) | $48M |
| YoY revenue growth | 62% |
| Enterprise revenue share | 68% of ARR |
| Self-serve revenue share | 22% of ARR |
| Top 20 enterprise accounts share | 41% of ARR |
| Additional Operating Data | Current State |
| --- | ---: |
| Gross logo retention, enterprise | 93% |
| Average enterprise sales cycle | 5.5 months |
| Self-serve trial-to-paid conversion | 3.8% |
| Monthly API developers signing up | 18,000 |
| Engineering capacity available for new initiatives | ~120 engineer-months over next 2 quarters |