Perception Deployment Engineer

Responsibilities

• Develop production-level, low-latency, memory-safe C++ and CUDA code for real-time perception algorithms on vehicle systems.
• Optimize multimodal sensor-fusion models, LLMs, VLMs, and other large-scale models using quantization, pruning, and mixed-precision inference.
• Architect model conversion and compilation pipelines using TensorRT for edge deployment.
• Perform model parity checking, accuracy recovery, and latency benchmarking between PyTorch models and compiled edge binaries.
• Develop and optimize custom ML operations and TensorRT plugins with efficient CUDA kernels for AI accelerators.

Requirements

• Production-level C++ and Python programming experience, including concurrent, memory-safe, real-time inference code for edge devices.
• Deep expertise in model compression, including PTQ, QAT, pruning, and mixed-precision inference such as INT8, FP8, BF16, and FP16.
• Experience optimizing large-scale multimodal sensor-fusion models, LLMs, VLMs, or VLAs using efficient attention and KV-cache optimization.
• Extensive experience with model conversion and compilation pipelines including ONNX, TensorRT, and torch.compile, along with latency benchmarking and model-quality parity evaluation.
• Proficiency in low-level AI accelerator programming, custom ML operations, TensorRT plugins, and CUDA kernel optimization.
• Familiarity with autonomous-driving perception algorithms, including temporal 3D object detection, BEV, and 3D Occupancy Networks, is a bonus.
• Experience with vision, LiDAR, and radar sensor processing, autonomous-driving VLM/VLA or foundation models, and TensorRT-LLM is a bonus.