Professional Experience

BIGO TECHNOLOGY PTE. LTD., Guangzhou, China
- Senior Engineer, Video Codec & Content Production Department 2020 - 2021
BIGO TECHNOLOGY PTE. LTD., Guangzhou, China
- Engineer, Video Codec & Content Production Department 2019 – 2020

Project Experience

Independently integrated ARKit into the client SDK, enabling rapid deployment of mainstream AR sticker features.
Responsible for developing the adaptation layer and functional extensions for the proprietary engine, enhancing the AR content production pipeline.

Independently resolved exposure logic issues across multiple scenarios based on photography theory, achieving accurate metering and significantly improving product experience.

Coordinated research on dehazing algorithms to improve image clarity on low-end devices and address color loss issues during transcoding.
Independently implemented and integrated image quality enhancement algorithms into the Android SDK.
Participated in low-level library development and NEON/SIMD instruction optimization to ensure efficient on-device algorithm performance.
Independently developed a camera lens dirt detection feature, resulting in a granted patent.

Conducted in-depth analysis of blocking artifacts in short video/live streaming transcoding, proposing innovative server-client joint solutions and driving issue resolution.
Independently optimized existing deblocking models and achieved efficient server deployment using OpenVINO.
Participated in the development of hybrid encoder (AV1, H264) deblocking algorithms, enabling on-device deployment.
Led the development of NEON and SIMD instruction set acceleration modules, significantly improving algorithm execution efficiency.

Led technical research and solution design for salient region detection algorithms, covering both traditional methods and deep learning models.
Designed and trained the SeeleNet model, achieving high-precision segmentation with low computational complexity; successfully deployed in the encoding team’s production environment.
Proficient in using TensorRT and OpenVINO for algorithm deployment, improving overall system performance.
Participated in low-level library implementation and production delivery, ensuring real-time performance and stability.

Conducted comprehensive research on mainstream SOTA image denoising algorithms (including traditional and deep learning methods), performed technical iteration analysis, and presented findings in departmental meetings.
Independently reproduced BM3D and related algorithms, delivering high-quality analysis reports and reproducible code.

Programming Languages: Proficient in C/C++, Python; Basic knowledge of Objective-C, Java
Deep Learning Frameworks: Proficient in PyTorch; Intermediate in RKNPU
Deep Learning Acceleration: Proficient in TensorRT, OpenVINO
Instruction Set Optimization: Intermediate in NEON
Domains: Intermediate in Computer Graphics; Basic in Audio/Video Codec
English Proficiency: IELTS 6.5 (L: 7, R: 7, W: 5.5, S: 5.5); GRE 328 (V: 162, Q: 166)