Zhenwei Hu Email & Phone Number

San Francisco, California, United States

• Developed a Scratchpad for the new Grammarly android app using Kotlin flows, Jetpack compose and Room database.• Used Gnar events and SQL queries to monitor and analyze user engagement with the Scratchpad, tracking daily usage by 200+ users.

• Conducted experiments on HINCare, a Heterogeneous Information Network (HIN)-based recommendation systems for elderly care, to explore the accuracy and performance of recommendation models, such as TransE, TransH and TransE-KG, achieving Hit Ratio (HR) > 90% and Discounted cumulative gain (NDCG) > 0.4.• Adapted Reinforcement learning-based Meta-path Selection (RMS) algorithm to investigate performance of meta-path-based algorithms.• Performed massive statistical analysis on Hong Kong Sheng Kung Hui Tseung Kwan O Aged Care Complex (SKH) and Christian Family Service Center (CFSC) elderly dataset with 1500+ users and 25000+ relations.• Presented research results to Hong Kong Innovation and Technology Commission officials with technical reports. • Supervised by Prof. Reynold C.K. Cheng

Beijing, China

• Built a program performance monitoring system and visualization dashboard using Python Dash with Plotly.• Contributed pull requests to an internal monitoring system on the company Gitlab by optimizing program runtime.

Shenzhen, Guangdong, China

• Reconstructed a Python backend song recommendation system, reducing deployment time from 2 hours to 5 min.• Maintained the reconstructed system by adding new features such as generating system reports, recording failure reports in MySQL databases and automating data processing using Crontab jobs.• Resolved memory issues of a Python song recommendation program, reducing memory usage from 20GB to less than 2GB.

Shenzhen, Guangdong, China

• Implemented a real-time multi-threaded market data receiving and forwarding system using C++ TCP and UDP socket with multi-threaded safe queue, heartbeat and reconnect mechanism. • Broadcasted real-time market data to internal hosts using ZMQ Publish-Subscribe pattern and EPGM protocol, achieving broadcast bandwidth >35MB/s.• Compressed market data using C++ ZSTD library and reduced broadcast bandwidth requirements by 60%.• Packed the data forwarding system written in C++ in a Python API with Pybind11 for other teams’ use.

Serve as a TA of ENGG1330 (Computer Programming I).

Master Of Science - Ms, Computer Science, 3.95

Activities and Societies: The 2022-2023 ICPC USA Mid-central Regional Contest Champion The 2023-2024 ICPC USA Mid-central Regional Contest.

Bachelor Of Engineering - Be, Computer Science, 3.84

Activities and Societies: HKU Competitive Programming TeamFirst Class Honor Dean's Honors List (2019, 2021) Minor in Mathematics

Summer School, Computer Science, 4.0

Course taken: CS61B Algorithms & Data Structures: A+ MATH55 Discrete Mathematics: A+