We are recruiting on behalf of a client for an Embedded Vision ML Engineer with expertise in TinyML and low-power computer vision to work on projects involving gesture recognition and real-time image processing on constrained devices. The ideal candidate has experience with ultra-low-power image sensors (e.g., Himax HM01B0 ) and deploying lightweight ML models using frameworks like TensorFlow Lite, QKeras, or Kendryte K210’s KPU .
Responsibilities include :
- Developing and optimizing TinyML-based gesture recognition and low-power computer vision models
- Working with ultra-low-power image sensors such as the Himax HM01B0
- Deploying and maintaining models on embedded platforms using frameworks like TensorFlow Lite, QKeras, or Kendryte K210’s KPU
- Implementing quantization-aware training and model compression techniques
- Handling real-time image processing under power and performance constraints
Qualifications :
Experience developing vision models on embedded platformsSkills in quantization-aware training, model compression, and deployment of lightweight ML modelsFamiliarity with ultra-low-power sensors and constrained hardware environmentsKnowledge of TinyML toolkits and frameworks such as TensorFlow Lite, QKeras, or Kendryte K210 SDKAbility to troubleshoot and optimize embedded vision pipelinesPreferred qualifications :
Experience with ultra-low-power image sensors and embedded camera hardwareUnderstanding of real-time embedded system constraints and optimization strategiesContributions to open-source TinyML or embedded vision projectsExperience with low-latency computer vision applications on constrained devicesAdditional details :
Seniority level : Mid-Senior levelEmployment type : Full-timeIndustry : Semiconductor ManufacturingJ-18808-Ljbffr