đź‘‹ Hi There, I'm
Dawei Liu
đź‘‹ About Me
I'm Dawei Liu, a first year master student in Computer Graphics and Game Technology (CGGT) at the University of Pennsylvania. I hold a Bachelor’s degree in Software Engineering from Northeastern University (NEU).
Currently, I’m an SDE Intern at TikTok, working on the Commerce Ads team, where I focus on recommendation infrastructure and performance optimization for large-scale machine learning serving systems. Previously, I interned at Amazon and JD.com, where I contributed to the distributed systems observability and AI platforms.
While my academic focus is in computer graphics—from auto-rigging to interactive systems—I also love digging into infra, systems performance, and ML deployment. I thrive at the intersection of graphics, AI, and systems engineering.
đź’Ľ Internship Highlights
🔹 TikTok – SDE Intern, Commerce Ads (2025)
At TikTok, I work on the recommendation infrastructure that powers Commerce Ads delivery. My contributions focus on building high-throughput, low-latency serving systems for ad ranking and model execution. This involves optimizing real-time inference, load balancing, and user experience protection under high QPS. I also participate in designing new ad products with infra that ensures scalability and fault tolerance at global scale.
🔹 Amazon – SDE Intern, Global Mile (2024)
During my internship at Amazon, I developed a custom Java Agent to extend OpenTelemetry’s tracing capabilities. By using ByteBuddy, I instrumented annotated methods for payload introspection and built a full-stack visualization platform—including timelines, tree plots, and fuzzy-search UI for end-to-end trace data. I also engineered a “Loosely Link” module to logically connect microservices via business IDs, enhancing traceability in distributed Lambda environments.
🔹 JD.com – SDE Intern, Algorithm Tools (2023)
At JD.com, I worked on platform engineering for internal AI tooling. I redesigned a resource management service using Kubernetes' Informer + observer pattern, reducing start-up time by 20x. I introduced GitOps + Argo Workflows for cloud-native CI/CD, built Helm charts for privatized deployments, and improved code modularity for activity page generation using AIGC pipelines. My work enabled faster and more maintainable delivery of algorithmic components.
🧠Currently Working On
- Scalable and low-latency recommendation infra at TikTok Commerce Ads
- Exploring graphics + AI interaction models
- Real-time rendering techniques and stylized shading in OpenGL
- Writing about leetcode, systems, tools, and design patterns
🛠️ Tech Stack
- Languages: Java, C++, TypeScript, Python, Go, SQL, Swift, GLSL
- Graphics: OpenGL, Unity, WebGL, Maya API, RigNet, ARKit
- Infra & Systems: OpenTelemetry, gRPC, Redis, Kafka, RocksDB, RabbitMQ
- DevOps & Cloud: Kubernetes, Docker, Helm, ArgoCD, GitOps, AWS (Lambda, S3, Kinesis, DynamoDB)
- Frontend: React, Vue, ECharts, Arco Design
- Tools: Git, Vim, VSCode, Mermaid, VuePress, LaTeX
💬 Let’s Connect
Whether you're into graphics, recommendation systems, AI infra, or creative engineering, I'd love to connect and chat. Thanks for stopping by!