Applied ML Engineer (AI)

TL;DR

Applied ML Engineer (AI): Own and streamline the research-to-production pipeline for speech models, turning research checkpoints into production models with an accent on release gates, evaluation rigor, and production serving performance. Focus on building reproducible training/evaluation workflows, packaging and deployment paths, and closing the feedback loop so the next model ships faster and more reliably under real traffic.

Location: Remote (USA)

Salary: $150K–$220K base (equity, bonus available)

Company

Deepgram provides real-time speech-to-text, text-to-speech, and voice agent infrastructure via production-grade APIs and self-hosted/on-prem software.

What you will do

• Own the research-to-production pipeline: define the repeatable path from working research results to deployed, monitored, scaled services.
• Partner with research scientists to productionize new models by translating experimental training/evaluation code into robust, reproducible, well-tested workflows.
• Build and extend tooling and abstractions for training, evaluation, packaging, and deployment with minimal friction and maximum reproducibility.
• Design model release gates with automated evaluation, regression detection, and quality/latency/throughput checks.
• Optimize production serving (efficient inference, batching, memory/latency tuning, profiling) to meet economic and performance targets at scale.
• Instrument production behavior and feed results back to research to accelerate iteration; establish consistent benchmarking/validation across dev-to-production.

Requirements

• Strong software engineering fundamentals with proficiency in Python and experience writing production-quality, well-tested ML code.
• Hands-on experience taking ML models from research/prototype to production at scale (training plus shipping and operating).
• Working understanding of the modern deep learning stack (e.g., PyTorch) and the realities of training, evaluating, and serving large models.
• Experience building ML pipelines and tooling (training orchestration, evaluation harnesses, model packaging, deployment, or CI/CD for models).
• Experience with inference optimization for production workloads (latency, throughput, batching, and resource efficiency) and comfort operating across distributed systems and GPU compute (cloud and/or bare metal).
• Experience with research-to-production handoff and automated evaluation/release-gating systems (regression detection across model versions).

Culture & Benefits

• AI-first mindset: actively use and experiment with advanced AI tools, and integrate AI into day-to-day work.
• Fast iteration: expect day-to-day work to evolve quickly as models and workflows improve.
• Builder role focused on measurable impact—what runs in production is the success metric.
• Compensation includes base salary plus equity and bonus (10% annual bonus mentioned).