Embedded ML Engineer Edge Accelerator

Role Overview: You are a passionate and skilled Embedded ML Engineer responsible for working on cutting-edge ML inference pipelines for low-power, real-time embedded platforms. Your role involves designing and deploying highly efficient ML models on custom hardware accelerators like Hailo, Coral (Edge TPU), Kendryte K210, and Torrent/BlackHole in real-world IoT systems. This position combines model optimization, embedded firmware development, and toolchain management, where you will be translating large ML models into efficient quantized versions, benchmarking them on custom hardware, and integrating them with embedded firmware pipelines interacting with real-world sensors and peripherals. Key Responsibilities: - Convert, quantize, and compile models built in TensorFlow, PyTorch, or ONNX to hardware-specific formats. - Work with compilers and deployment frameworks like TFLite, HailoRT, EdgeTPU Compiler, TVM, or ONNX Runtime. - Utilize techniques such as post-training quantization, pruning, distillation, and model slicing. - Integrate ML runtimes in C/C++ or Python into firmware stacks built on RTOS or embedded Linux. - Manage memory, DMA transfers, inference buffers, and timing loops for deterministic behavior. - Profile and optimize models for latency, memory usage, compute load, and power draw. - Collaborate with hardware and firmware teams to design unit, integration, and hardware-in-loop (HIL) tests. - Build reproducible benchmarking scripts and test data pipelines. Qualification Required: - Education: BE/B.Tech/M.Tech in Electronics, Embedded Systems, Computer Science, or related disciplines. - Experience: 24 years in embedded ML, edge AI, or firmware development with ML inference integration. - Technical Skills Required: Strong experience in C/C++, basic Python scripting, experience with RTOS or embedded Linux, knowledge of quantization-aware training, and model conversion pipelines. - Media & Sensor Stack: Ability to work with input/output streams from cameras, IMUs, microphones, etc. - Tooling & Debugging: Proficient in Git, Docker, cross-compilation toolchains, debugging with SWD/JTAG, GDB, or serial console-based logging. Additional Details: EURTH TECHTRONICS PVT LTD is a cutting-edge Electronics Product Design and Engineering firm specializing in embedded systems, IoT solutions, and high-performance hardware development. They provide end-to-end product development services and have expertise in embedded software, signal processing, AI-driven edge computing, RF communication, and ultra-low-power design. The company's vision is to advance technology and innovation in embedded product design, focusing on scalability, security, and efficiency to empower businesses with intelligent, connected, and future-ready solutions. Role Overview: You are a passionate and skilled Embedded ML Engineer responsible for working on cutting-edge ML inference pipelines for low-power, real-time embedded platforms. Your role involves designing and deploying highly efficient ML models on custom hardware accelerators like Hailo, Coral (Edge TPU), Kendryte K210, and Torrent/BlackHole in real-world IoT systems. This position combines model optimization, embedded firmware development, and toolchain management, where you will be translating large ML models into efficient quantized versions, benchmarking them on custom hardware, and integrating them with embedded firmware pipelines interacting with real-world sensors and peripherals. Key Responsibilities: - Convert, quantize, and compile models built in TensorFlow, PyTorch, or ONNX to hardware-specific formats. - Work with compilers and deployment frameworks like TFLite, HailoRT, EdgeTPU Compiler, TVM, or ONNX Runtime. - Utilize techniques such as post-training quantization, pruning, distillation, and model slicing. - Integrate ML runtimes in C/C++ or Python into firmware stacks built on RTOS or embedded Linux. - Manage memory, DMA transfers, inference buffers, and timing loops for deterministic behavior. - Profile and optimize models for latency, memory usage, compute load, and power draw. - Collaborate with hardware and firmware teams to design unit, integration, and hardware-in-loop (HIL) tests. - Build reproducible benchmarking scripts and test data pipelines. Qualification Required: - Education: BE/B.Tech/M.Tech in Electronics, Embedded Systems, Computer Science, or related disciplines. - Experience: 24 years in embedded ML, edge AI, or firmware development with ML inference integration. - Technical Skills Required: Strong experience in C/C++, basic Python scripting, experience with RTOS or embedded Linux, knowledge of quantization-aware training, and model conversion pipelines. - Media & Sensor Stack: Ability to work with input/output streams from cameras, IMUs, microphones, etc. - Tooling & Debugging: Proficient in Git, Docker, cross-compilation toolchains, debugging with SWD/JTAG, GDB, or serial console-based logging. Additional Details: EURTH TECHTRONICS PVT LTD