SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Deepgram is the leading platform for Voice AI, providing real-time APIs for speech-to-text, text-to-speech, and voice agents at scale. The company has processed over 50,000 years of audio and transcribed more than 1 trillion words, serving 200,000+ developers and 1,300+ organizations including Twilio, Cloudflare, and Vapi.
You will lead the design and implementation of Deepgram's internal data and ML training systems. This is a high-impact backend engineering role focused on building secure, robust, and scalable services that directly enable the company's AI research and product teams.
Key responsibilities include:
- Improve core data ingestion, cataloging, and transformation services, including networking, speech processing, audio transcoding, and performance optimization
- Engineer ML training code for performance and extensibility, enabling large-scale training of new foundational model architectures
- Develop processes for measuring and optimizing system performance
- Debug complex system issues spanning networking, scheduling, databases, and persistence layers
- Partner with DataOps and Research teams to design and implement end-to-end services and products
You should thrive in a fast-paced, AI-first environment where learning new skills on-the-fly is expected. The role requires balancing decisions about product maturity and when to make minimally invasive changes versus incorporating detailed design work.
Required qualifications: 3+ years of industry experience; strong programming in Rust (or C/C++) with Python competence; excellent communication and organizational skills; deep version control (git) and UNIX systems experience.
Desirable: experience with modern ML frameworks (PyTorch), knowledge of CNN/RNN/Transformer architectures, audio processing experience.
