影像挖掘算法工程师-地理位置中台
Core
Develop visual perception, multi-modal understanding, and change detection algorithms for complex road scenarios to automate road network data production using VLMs and large models.
Role type
Senior IC machine-learning engineer (computer vision & multi-modal)
Builds
Automated road network data production pipelines, road element topology reasoning, and change detection systems.
Domain
Autonomous driving / Computer Vision / Large Language Models
Deliverable
production ML models
Required skills
Deep learning, computer vision, Python, C++, PyTorch, object detection, semantic/instance segmentation, OCR, BEV perception, visual retrieval
Preferred skills
VLMs, visual foundation models, open-set object detection, natural language to road element parsing, VLA, world models, open-world perception, multi-modal fusion, active learning, data closed-loop
Technologies
DINOv2, DINOv3, CLIP, Qwen-VL, InternVL, LLaVA, Florence, Grounding DINO, PyTorch
Responsibilities
Research and implement BEV/3D road element modeling and topological reasoning; Develop human-in-the-loop algorithms for road change detection; Track and deploy transferable technologies from VLA and world model research.
Seniority
Senior, hands-on IC
