CV

You can find the PDF here.

Working Experience

MiTAC Information Systems Corp. – Apr.2024 ~ Present

Software Engineer; Company Homepage: https://www.mitaccomputing.com/; Newark, California

• Engineered a core service in Go with the Gin framework to provide a unified database access interface, featuring automated management of multiple databases, connection pools, and thread-safe transactions.
• Architected a multi-tiered service infrastructure with a reverse proxy to manage both internal and external API requests efficiently.
• Designed and developed a secure authentication service and API gateway in Go, implementing Role-Based Access Control (RBAC) for API routes using JWT.
• Developed robust RESTful API services in Python with FastAPI for CRUD operations on database entities and the generation of data reports.
• Successfully refactored and upgraded a legacy .NET client application to use Oracle.ManagedDataAccess, eliminating the dependency on Oracle Instant Client for end-users.
• Led the migration of internal codebases from TFS to GitLab, establishing a modern version control workflow and improving development efficiency.
• Established CI/CD pipelines to automate and streamline service deployments using Docker containers, orchestrated with Docker Compose for a declarative and version-controlled infrastructure.
• Re-engineered a Xamarin mobile application in Flutter for warehouse logistics management, successfully deployed it to over 100 handheld devices across multiple warehouses.
• Maintained and enhanced key internal systems, including the Shop Floor Control System (SFCS) and Logistic Information System (LIS), utilizing C# (.NET), ASP.NET, DevExpress, and PL/SQL.

Sichuan Hwadee Information Technology Ltd. – Jun.2020 ~ Jul.2020

Crawler Engineer Intern; Company Homepage: http://www.hwadee.com/; Chengdu, China

• Developed a web crawler using Scrapy to extract metadata for the top 1000 rated movies from IMDb.
• Utilized Redis to manage the queue of target URLs for the crawler.
• Automated the data pipeline by scripting the transfer of scraped data into a structured MySQL database for analysis.

UESTC Center for Future Media – Oct.2019 ~ Jun.2022

Deep Learning Research Assistant; Laboratory Web: https://cfm.uestc.edu.cn/; Advisor: Xing Xu; Chengdu, China

• Innovated a novel method for multi-source domain adaptation using partial feature alignment, implemented in PyTorch.
• Conducted extensive experiments to validate the model’s state-of-the-art performance, leading to a first-author publication at the prestigious CVPR conference.
• Authored the Statement of Claims for a patent based on the research, showcasing its novelty and potential for practical application.

Education

University of California, San Diego (UCSD) – Sep.2022 ~ Mar.2024

• Major: Computer Science | Cumulative GPA: 4.0/4.0
• Major Coursework: Principles of Computer Operating Systems, Recommender Systems, Probabilistic Reasoning & Learning, Convex Optimization, Search and Optimization, Graduate Networked Systems, Statistical Natural Language Processing, Compiler Construction, Parallel Computing, Data Science Programming, Principles of Computer Architecture, Mathematics for Robotics.

University of Electronic Science and Technology of China (UESTC) – Sep.2018 ~ Jun.2022

• Major: Computer Science and Technology | Cumulative GPA: 3.96/4.0
• Major Coursework: The C/C++ Programming, Data Structure and Algorithm, Computer Networks, Computer Operating System, Artificial Intelligence, Software Engineering, Principle and Application of Database, Principles of Computer Organization, Fundamentals of Compiler, Computer Architecture.

Publications

Yangye Fu, Ming Zhang, Xing Xu, Zuo Cao, Chao Ma, Yanli Ji, Kai Zuo, Huimin Lu; Partial Feature Selection and Alignment for Multi-Source Domain Adaptation; Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021, pp. 16654-16663

Research Experience

Partial Feature Selection and Alignment for Multi-Source Domain Adaptation – May.2020 ~ Jun.2021

• Unified issues of “partialness” to a more generalized Multi-Source Domain Adaption (MSDA) task as Multi-Source Partial Domain Adaption (MSPDA) and proposed a novel model termed Partial Feature Selection and Alignment (PFSA) to cope with both MSDA and MSPDA scenarios.
• Employed a feature selection vector based on the correlation among the features of multiple source domains and the target domain and designed three effective feature alignment losses to align the features by preserving the domain information of data sample clusters in the same category and discrimination between different classes.
• Extensive experiments conducted on three prevailing MSDA datasets exhibited that the model outperformed the state-of-the-art MSDA and PDA methods in both MSDA and MSPDA tasks.

Projects

Some Utility Tools

• Math renderer: write $\LaTeX$ equations in the browser and copy them to MS Word
• MIPS machine code generator: generate machine code from MIPS assembly
• AI Image Editor: a UI tool utilizing Gemini Image APIs (Nano Banana) for image generation and editing
• Renpy Translator: use LLM to translate Renpy game scripts

Lens Simulator – Oct.2024

• Project Github (Python): https://github.com/Fyy10/lens-simulator-py
• Simulate lens behavior through path tracing
• Vector manipulations & calculations with NumPy
• Experiments for lenses with different focal lengths, apertures, focal distances, etc.
• Results presented with Matplotlib

Cache Simulator – Nov.2023

• Implemented a double-layer LRU cache with C/C++
• Reimplemented the dedicated auxiliary data structures for the LRU cache such as Array, LinkedList, and HashSet
• The L1 cache (includes instruction cache and data cache) and the L2 cache are both set-associative
• Designed a next-line prefetcher for the cache and significantly reduced cache miss rate
• The double-layer LRU cache is tested through the cache configurations of MIPS R10K and Alpha 21264 on multiple program traces

FA Language – Aug.2023 ~ Nov.2023

• Project GitHub (C/C++): https://github.com/Fyy10/FAL
• A C-style compiler built with Flex, Bison, and LLVM
• Lexical analysis with Flex
• Semantic parsing with Bison
• Code generation and optimization with LLVM

Branch Predictor – Oct.2023

• The C++ implementation of multiple branch predictors proposed in different literatures
• Defined an abstract class Predictor as the interface where all implemented predictors are derived from this base class
• Reimplemented the GSharePredictor to follow the interface specification of Predictor
• Implemented the tournament predictor of the Alpha 21264 microprocessor
• Implemented the perceptron predictor proposed in the paper Dynamic Branch Prediction with Perceptrons
• The perceptron predictor was able to outperform the GSharePredictor by a large margin (4.6 less mis-predictions per 1k instructions)

Ray Tracer – Jul.2023 ~ Aug.2023

• Project GitHub (C/C++): https://github.com/Fyy10/ray-tracer
• Developed a ray tracer from scratch in C++, implementing fundamental computer graphics concepts.
• Engineered classes for 3D vectors, rays, and colors, and defined interfaces for diffuse, metallic, and dielectric materials to simulate realistic light interactions.
• Implemented advanced camera features including antialiasing and defocus blur to enhance image quality.

Snek Compiler – Apr.2023 ~ Jun.2023

• Project GitHub (Rust): https://github.com/Fyy10/snek-compiler
• Developed an x86_64 compiler for a Lisp-like language called ‘snek’ using Rust.
• Implemented a parser for the abstract syntax tree (AST) leveraging Rust’s pattern matching capabilities.
• Successfully implemented core language features including numerical and logical operators, variable bindings, conditional and loop statements, functions, and tuple-based structured data.
• Adhered to the System V ABI for function calling conventions.
• Managed heap-allocated data structures using Rust libraries, with low-level memory management in Assembly.
• Implemented a garbage collector using the LISP 2 mark-compact algorithm to manage memory automatically.

SurfStore: A Distributed and Decentralized File Synchronization System – Jan.2023 ~ Mar.2023

• Project GitHub (Go): https://github.com/Fyy10/SurfStore
• Engineered a distributed, decentralized file synchronization system from the ground up.
• Designed and implemented a key-value store for block-level data storage, with a local SQLite index for efficient client-side metadata lookups.
• Architected a scalable system that decouples metadata and content storage across distributed servers.
• Utilized gRPC and Protocol Buffers to define efficient, strongly-typed client-server communication protocols.
• Implemented a consistent hash ring for deterministic data placement and retrieval, ensuring balanced load distribution.
• Guaranteed system reliability and fault tolerance by integrating the Raft consensus algorithm for the metadata service.

Nachos 5.0j – Oct.2022 ~ Dec.2022

• Project Introduction (Java): https://cseweb.ucsd.edu/classes/fa22/cse120-a/projects/index.html
• Based on a Java implementation of a simulated operating system
• Completed the clock interrupt and thread join functionalities and implemented the rendezvous system call with condition variables and locks
• Developed demand paging and swapping with page fault handler and a swap file for supporting virtual memory and multiprocessing
• More on OS: A simulated shell and scheduler (C/C++)

File Backup System – Aug.2021 ~ Sept.2021

• Project GitHub (C/C++): https://github.com/Fyy10/File-Backup-System
• A file backup application with GUI developed in C/C++
• Led a group of 3, wrote the GUI with Qt, and implemented the backup-on-save feature with Linux inotify APIs
• The listening thread was created and managed by pthreads library

The Sieve of Prime Numbers and N-Body Simulation (Paralleled Computing) – Mar.2021 ~ May.2021

• Project GitHub (C): https://github.com/Fyy10/Parallel_Computing
• Implemented the Sieve of Eratosthenes Algorithm using MPI (Message Passing Interface).
• Computed the number of primes smaller or equal to the given number N using the MPI algorithm with three optimized approaches, which eventually achieved significant improvement in performance.
• Simulated the gravity interactions and the motions of 4096 bodies with CUDA C and reached more than 1500x speedup compared with sequential code on the target machine.

Applications and Challenges of Artificial Intelligence – Sept.2020 ~ Nov.2020

• Project GitHub (Python): https://github.com/Fyy10/Python-Playground/tree/master/AI
• Learned fundamental concepts and acquired skills in artificial intelligence including decision tree (ID3 algorithm), backpropagation in neural networks, fully connected networks for MNIST classification, LeNet for CIFAR10 classification, and VGG16 for flower classification.
• Led a team of 4 to attend the Kaggle Competition (Kannada-MNIST), where I managed to implement the designed models in Python utilizing PyTorch, and delivered the defense presentation.

Image Retrieval Engine – May.2020 ~ Jul.2020

• Project GitHub (Python): https://github.com/Fyy10/Python-Playground/tree/master/PCA
• Search for similar images based on principle component analysis (PCA)
• Data distribution analysis and visualization on CIFAR10 and DomainNet datasets

Convolutional Neural Networks for Visual Recognition – Jan.2020 ~ Feb.2020

• Assignment GitHub (Python): https://github.com/Fyy10/ML-DL_Practice/tree/master/cs231n
• Improved my programming skill and understanding of computer vision through finishing Stanford CS231n assignments that covered topics including K-nearest neighbor algorithm, fully-connected networks and convolutional networks, generative adversarial networks, and style transfer.

Artificial Intelligence Poet System – Dec.2019 ~ Feb.2020

• Project GitHub (Python): https://github.com/Fyy10/AI-Poet
• Constructed a model to automatically write Chinese poems.
• Implemented the model in Python with PyTorch as the deep learning framework.
• Applied NLP techniques including RNN, LSTM, CharRNN model, Seq2Seq model, and Attention Mechanism.
• Trained the model on a GTX 1050 Ti graphic card and released the trained model on GitHub.

Skills

• Languages: Python, Go, C#, PL/SQL, C/C++, Dart, Rust, Java, Intel Assembly
• Technologies: FastAPI, Gin, Docker, Docker Compose, Git, Flutter, PyTorch, NumPy, gRPC, MPI, CUDA, Qt, ASP.NET, DevExpress
• Databases: Oracle (PL/SQL), MySQL, SQLite, Redis
• Other: $\LaTeX$

Language

• Chinese
• English
• Japanese (a little)