• First and the only QA Engineer in the team. Established entire QA workflow from scratch, created and executed comprehensive Test Plan (manual and automated) for Mobile App, Tesla Energy Products and Vehicles. • Performed System Integration Tests between Tesla Energy devices and Android/iOS devices, focusing on cameras, Wi-Fi, and BLE-related functionalities. • Designed and deployed end-to-end Mobile App Automated Validation Pipelines on Cloud and HIL environment, integrating tools such as Xcode, Appium, DeviceFarm, Jenkins, and Xray, with detailed test reports and loggings. • Collaborated with PMs, designers, and developers on new product launch, ensuring excellent user experience. Innovated ML and image processing methods to automate complex test cases, enhancing testability with stakeholders • Significantly improved test coverage and operational efficiency, reducing manual regression test effort by over 70%

• Hosted defect triage meeting and organized resources to anticipate, plan, and solve issue with PMs and developers. • Designed and established Jenkins CI/CD Pipeline from scratch independently with Cypress and Selenium, generated over 300 automated API and Frontend test cases under 3 different network environments with high quality • Lead QA team, established end to end QA workflow from scratch and continuous optimization • Integrated AWX and Jenkins for better version control, validate system service robustness before deployment • Responsible and mentor junior engineers for end to end quality of major deliverables in 3 product lines

• Worked closely with different team and shared quality assurance status report including Development and Business • Troubleshoot and diagnosed issues including self-driving system, software, hardware, and network in real time manner, quickly determine root cause and provide technical solution, also forward and prioritize issues to stakeholders • Worked with AWX and Ansible playbooks for physical instance provisioning, configuration management, patching and software deployment • Automated routine tasks and testing with Shell scripting and Python

Trained SSD-MobileNet with SageMaker, deployed with 3x faster inference time on Nvidia Jetson Nano by TVM

• Trained License Plate Detection and Recognition (OCR) model on AWS EC2 instance and reached 98% accuracy

Project: Low-shot Object Detection Recall Improvement • Implemented metric learning loss on region proposal network to increase low-shot object recall, successfully reached proposal [email protected] to 0.89 by single-shot dataset

• Accomplished fine-tuning on YOLO model by only 500 new images to improve performance on specific scene with customer issues, successfully enhancing [email protected] from 0.263 to 0.394 and implemented in production