多模态算法工程师（自动驾驶云）-火山引擎
Core
Exploring large model applications in the automotive industry, including scene tagging, generation, generalization, behavior understanding, and image-text search; developing multi-modal Agent technologies for intelligent driving data processing.
Role type
Senior IC multi-modal algorithm engineer (autonomous driving cloud)
Builds
Data cleaning, model training/evaluation/fine-tuning, deployment optimization, and multi-modal data screening/preprocessing pipelines for the intelligent driving platform.
Domain
Automotive industry + Large Language Models (LLMs) + Multi-modal AI
Deliverable
production ML models
Required skills
Large model architecture (Transformer, Embedding, Multi-modal), Model optimization (LoRA, Quantization, Distillation), Deep learning frameworks (PyTorch, TensorFlow), Distributed training (DDP, FSDP), Agent components (Function Calling, RAG, Planning), Agent orchestration (LangChain, AutoGen), Python, C++
Preferred skills
Experience with DDP/FSDP, Strong debugging and profiling skills
Responsibilities
Collaborate with product teams to understand business needs and develop product solutions; Apply multi-modal Agent technology to automate data filtering, preprocessing, hard case mining, and dataset creation; Optimize model deployment and performance.
Seniority
Senior, hands-on IC
