CV
Education
National University of Singapore | Aug. 2024 – Jan. 2026 (expected)
Master of Computing (General Track) | Full-time | Singapore
- GPA 4.75/5, Relevant coursework includes: Computer Systems and Applications, Data Structures and Algorithms, Database Design and Programming, Parallel and Distributed Algorithms, Neural Networks and Deep Learning, etc.
Beijing University of Posts and Telecommunications | Sept. 2020 – Jul. 2024
BSc(Eng) Telecommunications Engineering with Management (Multimedia Track) | Full-time | Haidian, Beijing
- Grade 88/100 (Top 20%). Relevant coursework includes: Advanced Programming in C (96), Data Structures (91), Software Engineering (91), Internet Applications (92), Neural Networks and Deep Learning (92), etc.
Work Experience
Alibaba | May 2025 – August 2025
Taotian(Taobao and Tmall) Group - Alimama - Ad Technology - AI Inference Intern | Beijing
- Optimized Taotian’s internal video generation AIGC inference services for Taobao advertising creatives (淘宝星辰).
- Experimented with various cutting-edge sparsification techniques from academia, and proposed feasible solutions.
- Developed higher-performance operators based on the finalized sparsification scheme, and collaborated with NVIDIA on quantization acceleration.
- Achieved end-to-end acceleration ratios of more than 2x, and performance can be further improved when combined with technologies such as TeaCache.
- Gained exposure to other AI inference technologies, such as LLM serving.
Luchen Technology Co. Ltd. | Jan. 2025 – Apr. 2025
Luchentech (Colossal-AI) - HPC-AI Cloud - Cloud Platform Developer Intern | Singapore
- Full-cycle backend service development in Go, utilizing technologies like go-zero, PostgreSQL, and Kubernetes Helm.
- Enhanced the billing system to enable payment mode updates without cluster downtime.
- Implemented scheduled execution for elastic containers, enabling flexible resource management.
- Developed a sub-user system for resource and financial management with permission controls.
ByteDance Ltd. | Jun. 2023 – Jun. 2024
Seed - AI for Science - Backend Developer Intern | Haidian, Beijing
- Built a molecular structure-based database retrieval service (PostgreSQL+Flask+Serverless), enabling high-concurrency fuzzy searches and structured queries for molecular data.
- Standardized interfaces for molecular force fields and optimizers to facilitate modular reuse across services.
- Created a Serverless cluster parallel testing framework, leveraging compute pools in CI/CD to reduce unit test time by 75% and cut full release testing from days to hours.
- Led the design of a Kubernetes-based computational task scheduling system, supporting thousands of concurrent tasks with a non-technical-friendly configuration interface.
Academic Research
YOLO Model Inference Optimization for Mobile Devices | March 2025 – September 2025
National University of Singapore | Singapore
- Researched commonly used model inference acceleration techniques including various quantization methods, pruning, low-rank decomposition, and more.
- Validated quality and speed through adaptations and combination of different methods.
- (Planned) Deployment and testing on NVIDIA Jetson Nano.
Natural Language-Guided Neural Radiance Fields via Diffusion Models | Oct. 2023 – Jun. 2024
Beijing University of Posts and Telecommunications (Capstone Project) | Haidian, Beijing
- Developed a method combining Instruct-Pix2Pix and NeRF for natural language-based 3D scene editing with consistent rendering across different viewpoints.
- Evaluated the model using CLIP-based similarity scoring, demonstrating potential applications in virtual reality and real-time interactive environments.
Machine Learning Assisted Medical Research | Feb. 2023 – Aug. 2023
Beijing University of Posts and Telecommunications | Haidian, Beijing
- In collaboration with Peking University Third Hospital, our group utilized multimodal data, including medical records and EEG data, to predict the probability of delirium onset.
- Developed a machine learning algorithm using CNN, RNN, and Transformer architectures to predict delirium after hip fracture surgery in elderly patients. Delirium.
Awards
First Prize at National Finals | National University IoT Design Competition | Jun. 2021 – Sept. 2021
Deep Learning-based WIFI-CSI Sitting Detection System
- Utilized ResNet50 to detect body postures via Wi-Fi signals for non-intrusive, in-private health monitoring.
- Leveraged serverless cloud platforms for data inference, reducing computational load on terminal devices.
Additional Information
Extensive experience with Homelab and VPS, with strong network knowledge, and a spirit of discovery.
Development Projects: Personal LLM Chat Website, Java version of Wordle Wordle-Java, standards-compliant DNS server and client Project-DNS, assembly interpreter Assembly-VM, distributed high availability exercise CAST, etc.
Skills: C++, Python, Java, PostgreSQL, Git, PyTorch, Linux operation, Docker, Flask, Kubernetes, Nginx, etc.
Languages: English (professional working proficiency), Chinese (native proficiency).