AI Strategy: Scouting and incubating next-gen algorithms (LLMs, E2E Driving).
2. Cloud Architecture: Defining the data infrastructure that powers the "Data Closed-Loop"—from vehicle data ingestion and auto-labeling pipelines to cloud-based simulation and model training.
Key Responsibilities:
• Strategic Scouting (AI & Cloud Trends):
o Monitor global trends in Cloud-Native Automotive Architectures, focusing on Hybrid Cloud solutions (Edge-to-Cloud synergy) and Sovereign Cloud compliance in China.
o Scout emerging technologies in Cloud Simulation (e.g., World Models) and Data Management (Vector Databases for RAG, Synthetic Data Generation).
o Build the "AI & Cloud Tech Landscape Map," identifying opportunities to optimize computing costs while maximizing intelligence.
• Cloud & Data Architecture Design (The Backbone):
o Data Closed-Loop: Design the high-level architecture for automated data pipelines: Ingestion → Cleaning → Auto-Labeling (via Foundation Models) → Training → OTA.
o Hybrid Compute Strategy: Define what workloads run on the car (Edge NPU) vs. what workloads offload to the cloud (e.g., Shadow Mode data filtering logic vs. heavy model training).
o Simulation Infrastructure: Lead the technical definition for cloud-based large-scale simulation platforms needed for validating L3/L4 algorithms (ISO 8800 compliance).
• Deep-Dive Analysis (Feasibility & FinOps):
o Author "Deep-Dive Reports" on Cloud ROI & FinOps: Analyze the cost implications of scaling large model training and storage. Provide "Make vs. Buy" recommendations for cloud services (e.g., AWS vs. Azure vs. Private Cloud).
o Evaluate the feasibility of Vehicle-Cloud Collaborative Computing (e.g., RCP - Remote Control Protocol latency analysis).
• Incubation Leadership:
o Guide the Innovation Squad to build Cloud-Native PoCs (e.g., a RAG-based Knowledge Base hosted on cloud with vehicle connectivity, or a Shadow Mode data trigger mechanism).
o Ensure all cloud PoCs adhere to Cybersecurity and Data Privacy regulations.
Qualifications: