Open to opportunities

Zhang Wanyu

@zhangwanyu

Message

Machine learning engineer specializing in scalable backend systems and production ML deployments.

United States

Message

What I'm looking for

I seek roles where I can build and deploy production ML systems, work in agile teams, and drive measurable improvements in latency, reliability, and model performance.

I am a machine learning engineer with strong backend development experience and a track record of deploying production ML services. I combine full-stack engineering skills with applied ML to deliver scalable, low-latency systems.

At ByteDance and Tencent I built backend services, optimized query pipelines, and integrated caching and messaging systems to improve throughput and reliability. I increased test coverage and reduced latency and bug rates through disciplined engineering practices.

In research and product work I engineered large-scale geospatial ML pipelines and Graph Neural Network models, achieving high accuracy in crowd flow prediction and building simulated agent-based analyses for safety scenarios. I have also fine-tuned and deployed computer-vision models to edge devices with real-time performance.

I value agile collaboration, continuous delivery, and measurable impact—deploying containerized services, automating retraining and evaluation pipelines, and monitoring model performance in production to ensure robustness and scalability.

Experience

Work history, roles, and key accomplishments

Software Engineer

ByteDance

Feb 2025 - Jun 2025 (4 months)

Built backend services for short-video recommendation/search in Python + Flask, scaling to 500+ QPS at <200ms latency and reducing redundant calls by 25% via query optimization and Redis caching.

Python Flask Redis RESTful APIs Pytest Performance Optimization Microservices Linux

Software Development Engineer

Peking University

Mar 2024 - Jan 2025 (10 months)

Built large-scale geospatial ML pipelines processing 5TB+ road network data and 200K+ images; designed GNN + Ant Colony algorithms achieving 92% accuracy in crowd flow prediction.

Python Geospatial Ant Colony Algorithms Data Processing Simulation BIM

Machine Learning Engineer Intern

Microsoft

Jan 2023 - Jul 2023 (6 months)

Integrated NLP models into a Node.js backend for sentiment analysis improving efficiency 22% and built C++/OpenCV inference modules that reduced image API latency by 12%.

NLP Node C++OpenCV Docker Linux CI CD

Back End Developer

Tencent

Apr 2022 - Dec 2022 (8 months)

Processed high-volume weekly records with NumPy pipelines reducing manual analysis time 15%, refactored legacy code to cut MTTR from 4h to 2.5h, and integrated RabbitMQ to raise batch success rate to 94%.