7 months ago
Remote, United StatesMid Level
H1B sponsor
Base Salary
$160k - $220k/yr
Responsibilities
- Design and build CI/CD pipelines for ML model development, validation, and deployment.
- Architect model deployment pipelines spanning research, staging, and production environments.
- Build A/B testing infrastructure and controlled model rollouts.
- Implement monitoring for model accuracy, latency, drift, and regressions.
- Develop automated retraining pipelines triggered by data changes, performance degradation, or scheduled cadences.
- Create production-like build and test environments for research feedback.
- Establish model versioning, artifact management, and rollback capabilities.
- Collaborate with research engineers on model quality gates for production promotion.
- Build observability dashboards for model health across environments.
- Optimize model-serving infrastructure for latency, throughput, and cost efficiency.
Requirements
- 4+ years of experience in MLOps, DevOps, or infrastructure engineering focused on ML systems.
- Strong proficiency in Python and experience building ML workflow automation and tooling.
- Deep experience with CI/CD systems and software or model delivery pipelines.
- Hands-on experience with Docker and Kubernetes.
- Practical experience deploying and serving ML models in production.
- Familiarity with model evaluation, validation, quality assurance, monitoring, and observability for ML systems.
- Preferred experience with NVIDIA Triton Inference Server, TensorRT, or ONNX Runtime.
- Preferred background in speech, audio, or real-time media ML systems.
- Preferred experience with Terraform or Pulumi, and monitoring stacks such as Prometheus, Grafana, or Datadog.
- Preferred familiarity with GPU-accelerated inference optimization, feature stores, data versioning, ML metadata management, canary deployments, and progressive delivery.
Tech Stack
Categories
About Deepgram
Deepgram builds Voice AI infrastructure for developers, offering speech-to-text, text-to-speech, and voice agent tooling through real-time APIs and deployable on-prem software. Privately held and founded in 2015, the company is headquartered in San Francisco. Customers such as Twilio, Cloudflare, and Jack in the Box use its models to power production voice features; products include Nova-3 STT, Aura TTS, and a Voice Agent API.