Deepgram

2 weeks ago

Embedded AI Engineer, On-Device Models

✓ Verified Job

Canonical indexed version, validated from employer's careers page.

Source: Company careers pageValidated by: Fursa AI

Last checked: 22 hours agoStatus: Live

Job Summary

Build and optimize Deepgram’s speech and conversational models for on-device, real-time inference on resource-constrained hardware. You’ll define on-device inference architecture across diverse processors and accelerators, apply techniques like quantization and pruning, write performance-critical C/C++/Rust runtime code for embedded/RTOS environments, and integrate with edge runtimes and vendor NPU/DSP toolchains. You’ll also own benchmarking, OTA deployment pipelines, and work with silicon partners and research to keep models edge-friendly.

Location: California

Workplace: Remote

Employment Type: Full time

Job Function: Data Science & Machine Learning

Key Responsibilities

•Run Deepgram’s speech and conversational models on embedded and low-power consumer hardware by defining on-device, real-time inference architecture across processors and accelerators.
•Optimize models for constrained targets using techniques such as quantization, pruning, distillation, operator fusion, and architecture-specific compilation.
•Write and optimize performance-critical embedded runtime code in C, C++, and/or Rust, including bare-metal and RTOS environments (FreeRTOS, Zephyr).
•Integrate with edge inference runtimes and vendor NPU/DSP toolchains, mapping model graphs to on-device accelerators and CPU/GPU/NPU heterogeneity.
•Build on-device runtime plumbing (packaging, deployment pipelines, OTA updates, lightweight telemetry) and establish benchmarking/validation to prevent regressions.

Pay and Benefits

Salary: USD 219,300 - 274,100 annually

Equity and Bonus:Equity

Key Requirements

•Experience delivering production systems on resource-constrained hardware, including embedded systems, mobile, edge AI, or small low-power devices.
•Strong proficiency in C, C++, and/or Rust, with experience writing performance-critical code for constrained environments.
•Hands-on experience optimizing models for on-device deployment (e.g., quantization, pruning, knowledge distillation, or architecture-specific compilation).
•Familiarity with edge inference runtimes (ONNX Runtime, TensorRT, TFLite, ExecuTorch) and/or vendor-specific NPU/DSP toolchains.
•Experience working close to the metal in bare-metal/RTOS (FreeRTOS, Zephyr), embedded Linux, or microcontroller/edge SoC development.

Experience:Voice AIEmbedded systemsEdge AIOn-device inference

Skills:CommunicationBuilder mindset

Tech Stack:CC++RustFreeRTOSZephyrONNX RuntimeTensorRTTFLiteExecuTorchQuantizationPruningKnowledge distillationOperator fusionModel compilationOver-the-air updates

Company Brief

Deepgram

Deepgram builds a real-time Voice AI platform delivering speech-to-text, text-to-speech, and voice-agent APIs for developers and enterprises, focusing on low-latency, high-accuracy voice models and scalable deployment options.

Industry: API Platforms

Company Size: Medium (51 to 250 employees)

Growth: Scaleup

Valuation: Unicorn (USD 1B+)

Funding: Series C

Headquarters: San Francisco, United States

Founded: 2015

Glassdoor: 4.5

Website LinkedIn Glassdoor

Loading job details...

Preparing the role view and application actions.

Deepgram

2 weeks ago

Embedded AI Engineer, On-Device Models

✓ Verified Job

Canonical indexed version, validated from employer's careers page.

Source: Company careers pageValidated by: Fursa AI

Last checked: 22 hours agoStatus: Live

Job Summary

Location: California

Workplace: Remote

Employment Type: Full time

Job Function: Data Science & Machine Learning

Key Responsibilities

•Run Deepgram’s speech and conversational models on embedded and low-power consumer hardware by defining on-device, real-time inference architecture across processors and accelerators.
•Optimize models for constrained targets using techniques such as quantization, pruning, distillation, operator fusion, and architecture-specific compilation.
•Write and optimize performance-critical embedded runtime code in C, C++, and/or Rust, including bare-metal and RTOS environments (FreeRTOS, Zephyr).
•Integrate with edge inference runtimes and vendor NPU/DSP toolchains, mapping model graphs to on-device accelerators and CPU/GPU/NPU heterogeneity.
•Build on-device runtime plumbing (packaging, deployment pipelines, OTA updates, lightweight telemetry) and establish benchmarking/validation to prevent regressions.

Pay and Benefits

Salary: USD 219,300 - 274,100 annually

Equity and Bonus:Equity

Key Requirements

•Experience delivering production systems on resource-constrained hardware, including embedded systems, mobile, edge AI, or small low-power devices.
•Strong proficiency in C, C++, and/or Rust, with experience writing performance-critical code for constrained environments.
•Hands-on experience optimizing models for on-device deployment (e.g., quantization, pruning, knowledge distillation, or architecture-specific compilation).
•Familiarity with edge inference runtimes (ONNX Runtime, TensorRT, TFLite, ExecuTorch) and/or vendor-specific NPU/DSP toolchains.
•Experience working close to the metal in bare-metal/RTOS (FreeRTOS, Zephyr), embedded Linux, or microcontroller/edge SoC development.

Experience:Voice AIEmbedded systemsEdge AIOn-device inference

Skills:CommunicationBuilder mindset

Tech Stack:CC++RustFreeRTOSZephyrONNX RuntimeTensorRTTFLiteExecuTorchQuantizationPruningKnowledge distillationOperator fusionModel compilationOver-the-air updates

Company Brief

Deepgram

Industry: API Platforms

Company Size: Medium (51 to 250 employees)

Growth: Scaleup

Valuation: Unicorn (USD 1B+)

Funding: Series C

Headquarters: San Francisco, United States

Founded: 2015

Glassdoor: 4.5

Website LinkedIn Glassdoor