👋 Hi There, I'm

Dawei Liu

About Me

I’m Dawei Liu, a Master’s student in Computer and Information Science (CIS) at the University of Pennsylvania (Class of 2026), with a B.E. in Software Engineering from Northeastern University.

Currently, I’m an SDE Intern at TikTok (Shop Ads), working on Ads delivery, recommendation infrastructure, and performance optimization for large-scale distributed systems. Previously, I interned at Amazon and JD.com, contributing to observability frameworks, AI pipelines, and cloud-native deployment platforms. My work has been recognized with 2026 full-time return offers from both TikTok and Amazon.

I’m passionate about building scalable, efficient, and intelligent systems at the intersection of AI infrastructure and engineering excellence, solving complex problems that push the boundaries of system performance and reliability.

🎓 Education

🔹 University of Pennsylvania – M.S.E. in Computer and Information Science

Aug 27, 2024 – May 18, 2026

GPA: 3.88 / 4.00
Awards:
- Hagan International Scholarship

🔹 Northeastern University – B.E. in Software Engineering

Sep 01, 2020 – Jul 01, 2024

GPA: 3.95 / 4.00
Awards:
- National Scholarship
- Merit-based Scholarship
- Outstanding Student Leader
- Outstanding Graduate

🔹 Zhengzhou foreign language School

Sep 01, 2017 – Jul 01, 2020

💼 Professional Experience

🔹 TikTok – SWE Intern, Commerce Ads

May 19, 2025 - Dec 05, 2025

At TikTok, I focused on building scalable and fault-tolerant recommendation infrastructure for Commerce Ads. I designed and optimized systems to support high-throughput, low-latency Ads delivery, including a debugging and verfication framework for new ad formats, cache restructuring that improved stability under heavy traffic, and real-time Flink features for ranking and creative selection. My work directly enhanced the platform’s resilience, efficiency, and rollout velocity at global scale.

Built the Shop Ads Image Selection system, modeling image posterior features, applying exploration–exploitation ranking, and integrating multimodal LLM–based image quality evaluation, to serve high-performing images, driving a 20–30% advv uplift.
Built a multi-strategy AIGC image generation pipeline, auto-triggering updates for incremental SPUs via delivery stream and running daily scheduling to improve the top 90% cost-contributing SPUs. Synced results to TBase and index service via Flink.
Built Modular Preview Flow, a framework enabling stage-level entity injection (Ad, Creative, SPU, Image) across all delivery funnel stages with unified filtering log, improving debugging and verification efficiency for new ad formats and region rollout.
Enhanced product handler stability by redesigning cache mechanism and applying async batch fetching with Folly Future, mitigating 80% of failure spikes in high-traffic scenarios. Migrated online product value call to offline, reducing 13% latency.

🔹 Amazon – SDE Intern, Global Mile

Jun 11, 2024 - Sep 06, 2024

At Amazon, I developed a custom Java Agent to extend OpenTelemetry’s tracing, enabling end-to-end observability across distributed microservices and Lambda environments. I built full-stack tools for trace visualization and implemented a Loosely Link module that logically connected services without direct invocations. By improving traceability, reliability, and developer tooling, I enhanced the debugging and monitoring experience for complex distributed systems.

Developed a Java Agent that extends OpenTelemetry. Leveraged ByteBuddy to enhance methods annotated with @WithSpan, @Input, and @Output, enabling automatic tracing and payload collection.
Implemented frontend UI and backend APIs, with filters, aggregation, fuzzy search, pagination, and trace details visualizations including tree plot, table, timeline, end-to-end hyper process and span payloads (inputs, outputs and errors).
Enhanced end-to-end traceability with Loosely Link module that dynamically connects relevant services using business IDs and timestamps, enabling logical linkage even without direct calls (e.g., async MQ or partial service onboarding).
Utilized reflection to detect Lambda environments, and employed a lightweight Kinesis SDK. Implemented connection pooling, retry and flush sync mechanisms to ensure reliable data streaming within Lambda size constraints.

🔹 JD – SDE Intern, Algorithm Tools

Jul 24, 2023 - Oct 24, 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.

Refactored the resource management service with ConfigMap for automated and configurable resource-splitting solutions.
Utilized the Kubernetes Informer mechanism, observer pattern, async processing, row locking, and discard policies, to improve platform resource recalculation efficiency and achieve a 20x faster service startup speed.
Enhanced the CI/CD toolchain with cloud-native CI workflows using Argo Workflows and GitOps to trigger Argo CD.
Developed Helm charts to support robust on-premise and client-tailored deployments, enabling flexible private cloud delivery.
Refactored the campaign page generation service with the strategy pattern, improving modularity and reusability.
Designed generation pipelines integrated with AIGC services for automated creation of campaign page sections.

🛠️ Tech Stack

Language: Java, C/C++, Go, Python, JavaScript/TypeScript, HTML/CSS, Swift, SQL
Backend: Spring Framework, Guice, Coral, Thrift, Protobuf, MyBatis, MySQL, Redis, Kafka, RabbitMQ, ElasticSearch, OpenTelemetry
Frontend: React, Vue, Vite, ECharts, AWS UI, Arco Design; iOS: Swift, SwiftUI, ARKit
AI/ML: PyTorch, LLM (Transformer, RLHF, Token Pruning, CLIP), RecSys (ItemCF, Two-/Three-Tower, MTL)
Graphics: OpenGL, GLSL, Unity, Qt, Maya API (Plugin Development), Auto-Rigging (RigNet)
DevOps: Unix/Linux, Docker, Kubernetes, Helm, AWS (DynamoDB, Kinesis, S3, Load Balancer, CodePipeline, CDK)
Tools: Git, Vim, SSH, CI/CD, Shell, Markdown, LaTeX, Mermaid, VuePress

💬 Let’s Connect

Whether you're into Ads Delivery, recommendation systems, AI infra, or excellent engineering, I'd love to connect and chat. Thanks for stopping by!