Embedded AI Engineer (Android Automotive, On-Device Intelligence)
Posted Updated
- We are building on-device intelligence for a next-generation Android Automotive platform. This role owns the end-to-end lifecycle of embedded ML systems, ensuring models behave predictably and safely in production environments under real-world constraints such as latency, thermal limits, and functional safety
- Deploy and run production-grade ML inference and learning systems on Android Automotive (AAOS)
- Implement on-device multimodal LLMs, including schema design and safe dispatch to local vehicle APIs
- Integrate models using TensorFlow Lite, ONNX Runtime, or specialized vendor SDKs
- Profile and optimize models for strict latency, memory, power, and thermal budgets
- Instrument runtime performance across CPU, GPU, and NPU acceleration layers
- Design safety boundaries and guardrails for model outputs, including tool-call allowlists and fallback logic
- Interface directly with vehicle signals, sensors, and system services using C++ and JNI
- 3+ years of experience shipping ML inference on embedded, mobile, or automotive platforms
- Deep understanding of edge constraints including real-time behavior and memory pressure
- Strong proficiency in C++ and experience with native Android integration (JNI)
- BS, MS, or PhD in Computer Science, Electrical Engineering, or a related technical field
- Expertise in model optimization techniques such as quantization, pruning, and compilation
- Hands-on experience with Android system services or Android Automotive OS (AAOS)
- Experience integrating LLM function calling or tool execution with structured outputs
- Don’t meet every single requirement? If you’re excited about this role but your past experience doesn’t align perfectly with every qualification in the job description, we encourage you to apply anyway. You may be just the right candidate for this or other roles
- Background in running quantized LLMs on-device using llama.cpp or TFLite transformers
- Experience with Snapdragon Automotive, ARM Ethos, or specialized NPU pipelines
- Experience bridging cloud-trained models to resource-constrained embedded runtimes
- Familiarity with functional safety concepts (ISO 26262), sandboxing, or policy enforcement