About the RoleAs an Embedded AI Engineer, you will work closely with the AI research team to bring AI to Hark's next-gen hardware. You will be responsible for the full AI stack on the device, including data ingestion, model development, optimization, and deployment on embedded devices. You should have deep understanding of the constraints of an embedded system (compute, memory, power etc) and leverage your expertise in both embedded system software development and AI model deployment to deliver production-ready ML solutions on hardware
Responsibilities- Build data collection and ingestion pipelines for an embedded system including various sensors, at scale
- Work closely with model teams to co-design model architectures that meets the required latency, memory, power, and bandwidth
- Work with platform vendors to bring up toolchains, SDKs and new accelerator to ensure efficient model deployment and optimization
- Integrate ML inference into embedded firmware written in C, C++, or Rust
- Profile and optimize memory usage, power consumption, and real-time performance
- Evaluate and select silicon platforms (GPUs, NPUs etc.) for Hark's next gen on-device and edge deployment of a wide range of models
Requirements- 5 years of experience in machine learning engineering, with at least 2 years focused on embedded or edge ML
- Familiarity with embedded systems, and CPU/DSP/NPU HW architectures
- Hands-on experience with IMUs and other sensor types including accelerometers, gyroscopes, and microphones
- Experience building sensor data collection pipelines
- Familiarity and experience with embedded ML run times (e.g. TFLite, llamacpp, QNN)
- Experience optimizing models for deployment on microcontrollers and edge processors such as ARM Cortex-M/A, RISC-V, and DSPs
- Experience deploying workloads on NPUs or specialized accelerators for embedded systems
Bonus Qualifications- Experience with Audio/Voice/Vision models
- Experience with light weight LLM models
- Understand the performance characteristics of edge AI models, including CNN, RNN, transformers, KV-cache behavior, and their memory bandwidth requirements.
- Experience designing hybrid edge-LLM pipelines or integrating small language models on device
- Prior work on products in wearables, robotics, industrial sensing, or IoT
CompensationThe US base salary range for this full-time position is between $200,000 - $450,000 annually.
The pay offered for this position may vary based on several individual factors, including job-related knowledge, skills, and experience. The total compensation package may also include additional components/benefits depending on the specific role. This information will be shared if an employment offer is extended.
