Research Engineer, Post-Training

Deepgram • ,

6d • $150,000 - $250,000

About The Position

Deepgram is seeking a Research Engineer, Post-Training to improve its voice AI models by identifying customer problems, measuring model limitations, and delivering fixes to production. The role involves turning broad customer needs into clear metrics, finding effective approaches to address them, and shipping improvements for customers to use within weeks.

Projects may include expanding models to new languages and regions; improving text-to-speech pronunciation of names, terms, and numbers; separating speakers from background noise and crosstalk; providing early access to new capabilities such as steerable and cloned voices and using feedback to guide research; and releasing benchmarks for Voice AI use cases.

The engineer will identify model shortcomings and set improvement targets, design evaluations, curate or generate data, fine-tune models, build working systems, and ship them to production. Deepgram does not require a specific degree and invites candidates to demonstrate their abilities through challenging past projects, shipped systems, or open-source work. The company expects an AI-first approach, active use and experimentation with advanced AI tools, and adaptability in a rapidly changing work environment.

Requirements

• Deep experience training, fine-tuning, and evaluating deep learning models.
• Strong Python and PyTorch skills.
• A track record of shipping machine learning systems to production.
• Ability to own problems end to end and learn what is needed along the way.
• Good judgment about which problems matter and when a result is ready to ship.
• Comfort adopting new models and modes quickly, integrating AI into workflows, experimenting, and adapting to change.

Nice To Haves

• Experience with speech-to-text, text-to-speech, or voice agents.
• Experience with distributed training.
• Published papers, open-source contributions, or public benchmarks.

Responsibilities

• Identify where models fall short and set targets for improvement.
• Design evaluations to determine whether a model has improved and explain why.
• Curate and generate data and fine-tune models.
• Build working systems quickly and ship them to production.
• Own customer problems through to measurable improvement in production.
• Find effective and timely approaches to resolving customer challenges.
• Determine when results are ready to ship to customers.

What This Job Offers

• Job Type: Full-time
• Education Level: No Education Listed