Song Z.

Key Skills

Software

CVAT

Top Subject Matter

Automotive & ADAS Technology

Embedded Systems & Automotive Software

Data Validation & Automotive Telematics

Top Data Types

Computer Code Programming

Text

Document

Top Task Types

Computer Programming/Coding

Freelancer Overview

As an Automotive Embedded & AUTOSAR Development Engineer with 8+ years of hands-on experience in automotive MCU/ECU software development, I bring deep expertise in structured automotive data processing, system-level data validation, and quality-focused execution—core skills for high-quality AI training data preparation. My work spans CAN/Ethernet bus data analysis, ADAS module data validation, and diagnostic telematics data structuring, with a proven track record of ensuring data accuracy, consistency, and compliance with automotive functional safety standards (ISO 26262). Leveraging my background in embedded systems, UDS diagnostics, and automated testing, I design robust workflows for data labeling, cleaning, and annotation tailored to the rigorous requirements of automotive AI training pipelines. My education includes a Bachelor of Electronic Information Engineering from Chongqing Jiaotong University (2017), providing a strong foundation in signal processing and embedded data systems

Labeling Experience

ADAS Sensor Data Annotation & Validation (TC397 mADC Platform)

3D SensorSegmentation

Supported the Changan C2L ADAS project by annotating and validating multi-modal sensor and CAN bus data streams from the TC397-based mADC module. Key tasks included: 1. Annotating object detection targets (vehicles, pedestrians, lane markers, traffic signs) from raw camera/radar data using bounding boxes and semantic segmentation. 2. Cross-validating labeled sensor data against CAN bus signals (speed, brake status) to ensure spatiotemporal alignment, eliminating annotation errors. 3. Cleaning noisy data, filtering invalid samples, and organizing datasets into standardized formats for perception model training. Delivered 120,000+ validated samples with 99.5% labeling accuracy, adhering to ISO 26262 safety standards.

2024 - Present

Education

Work History