About the role
Deepgram is a leader in Voice AI, providing real-time speech-to-text and text-to-speech APIs used by over 200,000 developers and numerous organizations. The company focuses on building scalable, accurate, and cost-efficient voice models accessible via cloud or on-premises solutions. Their models process vast amounts of audio data, establishing a strong foundation in voice technology and AI-driven speech recognition.
The Applied ML Engineer role is part of the engineering team, focusing on adapting Deepgram's speech models for deployment on various edge and non-NVIDIA hardware platforms. The engineer will modify models to fit different hardware constraints, optimize performance, and validate accuracy and latency on real devices. The work involves collaboration with embedded AI engineers and platform vendors to ensure seamless deployment.
Candidates should have hands-on experience deploying machine learning models to edge hardware, knowledge of quantization and precision tradeoffs, and familiarity with inference runtimes like TFLite or OpenVINO. The role is remote, full-time, and suited for senior or staff-level engineers who enjoy solving complex deployment challenges and working at the intersection of hardware and AI software. The position emphasizes practical deployment over research, with a focus on delivering reliable, optimized models to customers.