Senior Embedded AI Researcher

DESCRIPTION

We are looking for a highly skilled and motivated engineer to join our team at the intersection of AI research and embedded systems. In this role, you will be responsible for bringing our cutting-edge computer vision models to life on our Ambarella CVxx-based dash cams. You will work closely with our research engineers to optimize and deploy neural networks for tasks such as object detection, video reconstruction, and collision prediction, all within the resource-constrained environment of an embedded system. This is a unique opportunity to have a major impact on the performance and capabilities of our products, and to help shape the future of AI in the automotive industry.

You are a good fit if you:

• Get excited about squeezing 10% more performance from a carefully tuned operator

• Can explain both ResNet architecture and L2 cache eviction policies

• Think about problems from both "what model architecture works best?" and "can this fit in 512MB of RAM?"

• Enjoy working at the intersection of research innovation and engineering pragmatism

• Are energised by the constraints of real-time, power-limited embedded systems

RESPONSIBILITIES

• Develop, optimize, and deploy deep learning models on our Ambarella CVFlow-based camera platforms , or future replacements in the future.

• Bridge the gap between our research and systems engineering teams, translating state-of-the-art AI models into highly efficient, production-ready code.

• Dive deep into the Ambarella CVFlow architecture to maximize the performance of our neural networks, leveraging your expertise in computer vision, hardware acceleration, and low-level programming.

• Analyze and profile model performance, identifying and addressing bottlenecks related to I/O, memory, and compute.

• Implement and experiment with model optimization techniques such as quantization, pruning, knowledge distillation strategies and network architecture search to meet strict latency and power budgets, while managing precision / recall and accuracy targets.

• Develop custom operators and layers optimized for CVflow's vision processing engine.

• Work with our systems engineers to design, implement and ultimately ensure the seamless integration of AI models with camera firmware and software.

• Stay up-to-date with the latest advancements in embedded AI, GenAI, and computer vision.

REQUIREMENTS

Technical skills & qualifications

• BS/MS/PhD in Computer Science, Electrical Engineering, or related field

• 3+ years of industry experience in embedded ML, edge AI, or related domains

• Track record of shipping ML models in production embedded systems

• Proficiency in C/C++ for embedded systems and performance-critical code.

• Understanding of computer architecture: cache hierarchies, memory systems, SIMD/vector processing.

• Knowledge of DMA, interrupts, and hardware-software interfaces.

• Hands-on experience with embedded AI platforms and hardware accelerators (e.g., Ambarella CVFlow, NVIDIA Jetson, Qualcomm Hexagon).

• Deep understanding of computer architecture, including concepts like kernel scheduling, I/O, memory management, and GPU/NPU pipelines.

• Experience with heterogeneous computing (CPU/GPU/NPU/DSP coordination).

• Prior experience deploying neural networks to resource-constrained devices (mobile, automotive, IoT)

• Professional experience with deep learning frameworks such as TensorFlow, PyTorch, or Caffe.

Bonus Points

• Direct experience with Ambarella SoCs and CVflow architecture

• Background in computer vision applications (object detection, segmentation, tracking)

• Experience with automotive ADAS development and safety standards

• Knowledge of video codec integration (H.264/H.265)

• Understanding of fixed-point arithmetic and numerical stability

• Experience with TensorRT, LiteRT, ONNX Runtime, or similar inference engines

• Contributions to ML deployment frameworks or toolchains

WHAT WE OFFER

• The chance to have a meaningful and positive impact on roads and cities’ safety

• Flexibility - in schedule, work location, benefits

• Competitive compensation

• A fast-paced work environment full of kind, smart, and highly driven people

• Great professional and personal growth opportunities

• Regular team-building events and well-being initiatives/facilities

WHO WE ARE

Nexar is building the real-world intelligence platform powering the physical AI era. As machines move into the real world—from autonomous vehicles to robotics—they all face the same challenge: they cannot scale or operate safely without access to high-quality, real-world data. Nexar solves this by operating one of the largest naturalistic driving intelligence platforms globally, powered by a live network of hundreds of thousands of cameras and a rapidly growing dataset of safety-critical events.

Our platform transforms this data into actionable intelligence used across multiple industries—from AV and ADAS to fleets, insurance, and public safety. By connecting real-world data supply with cross-industry demand, we enable organizations to better understand, benchmark, and improve how their systems perform in the real world. Every interaction strengthens the platform, making it smarter, more scalable, and increasingly valuable.

We’re building more than a product—we’re building the intelligence layer for the physical world. A platform that compounds with every mile driven, every dataset delivered, and every customer served.