Autonomous Driving

AUTONOMOUS DRIVING

I have led cloud infrastructure management across AWS, and private clouds, utilizing Terraform for Infrastructure as Code (IaC) and DevOps practices to build scalable, resilient systems. This includes optimizing compute, storage, networking, and security for big data workloads, ensuring high availability and compliance in regulated automotive settings.

🏆 UN R171.01 — Global ADAS Certification (March 2026)

In March 2026, Lotus Tech became the 2nd automaker globally—and the first with a China-built model—to achieve UN R171.01 certification from UNECE, the world's first harmonized technical regulation for Level 2 Driver Control Assistance Systems (DCAS). As the cloud and data platform architect, I built the infrastructure backbone that enabled this milestone: the ADAS data-closed-loop system for multi-modal sensor data processing, automated ML model training and evaluation, scenario extraction and edge case prioritization, and compliance documentation workflows meeting the regulation's four core pillars (functional performance, dynamic control, system boundaries, and human-machine interaction).

Autonomous Driving Toolchain Development

I led the deployment of the cloud-based toolchain for data-driven autonomous vehicle development on AWS Kubernetes clusters, integrating with S3 and private cloud storage for secure, privacy-compliant storage of converted vehicle data (ROS format). This platform manages multi-modal sensor data from LiDAR, cameras, and radar through end-to-end workflows, including collection, anonymization, processing, ML model training, ADAS function analysis, and planning & control algorithm simulation, while supporting efficient scenario extraction (RAG-based) and edge case prioritization to accelerate algorithm improvements. I ensured traceability, data quality governance, and adherence to automotive standards like E-NCAP. The secure pipelines enables rapid iteration on safety-critical features—optimized for cost-efficiency based on GPU utilization.

🎤 TechAD Europe 2025 Speaker
Presented groundbreaking work on autonomous driving toolchain development at TechAD Europe 2025 in Berlin

SaaS Platform Migration & Fleet Management

I led the migration of a SaaS platform for real-time consumer fleet data collection from Alibaba Cloud to AWS, re-architecting Kubernetes-based Java microservices with Nacos for service discovery. It features IoT Core with MQTT5 for telemetry ingestion, bidirectional communication with 2-way mTLS auth, Lambda for decoding, S3 for storage, and Glue/Redshift for analytics, processing streams like periodic signals, EDR uploads for AEB/SRS events, and syslog data. I implemented ingestion pipelines for hourly vehicle data cross-cloud synchronization, secure S3 uploads via temporary STS credentials and AKSK tokens, and real-time streaming through MSK (Kafka) and Kinesis Firehose for data analysis service.