Customer Service Machine Learning

EDUCATION

Yuqi Dai

Tel: +1-858-***-**** Email: ******@****.***

University of California San Diego (UCSD) San Diego, America MS in Computer Science and engineering Sept. 2024 - Dec 2025(expect) Ø GPA: 4.0 / 4.0

Huazhong University of Science and Technology (HUST) Wuhan, China B. Eng. in Computer Science and Technology Sept. 2020 - June 2024 Ø GPA: 3.86 / 4.0 Honors & Awards: Excellent Student Leader (2021 & 2022, Top 5%) INTERNSHIP

Convoloo C#, SQL, HTML, CSS, Web development Jan.2025 - Mar 2025, San Diego Software development engineer

Ø Designed and implemented a teaching and testing platform that supports multi-user access and real-time data updates. Ø Developed a dynamic website using the ASP.NET MVC framework, effectively employing the MVC architecture to separate business logic from the user interface, created robust controllers and intuitive views, leveraging HTML and CSS to construct a responsive web interface that adapts seamlessly to various devices. Ø Designed a database based on SQL Server and wrote SQL queries to support the data requirements of the application. Baidu, Intelligent Customer Service Team Transformer, Python, natural language processing Oct.2023 - Feb 2024, Beijing Machine Learning engineer

Ø Collaborated within a team to develop an intelligent customer service system based on a large language model. Implemented a multi-label text classifier using transformer, which is a key component of this intelligent customer service system. Applied data augmentation techniques to achieve an accuracy of over 0.9 for the classifier. Ø Developed a multi-agent collaborative dialogue system that utilizes a large model to efficiently generate text data for different scenarios, which is used for training, testing, and fine-tuning the language model. Ø Developed a website based on the Streamlit framework for uploading, downloading, and managing model training datasets. Ø Investigated LLM application frameworks like LangChain, AutoGen, etc. Delivered a presentation to the whole team. Yantu Technology, Machine learning team. Pytorch, Python, natural language processing Apr. 2023 - July 2023, Wuhan Machine Learning engineer

Ø Participated in the development of an intelligent chatbot based on Springboot, utilizing pre-trained language models to implement natural language processing functions. The chatbot is capable of understanding user intent and generating intelligent responses in real-time. My main responsibilities included backend development and error case analysis. Ø Implemented language models based on transformer decoders and deep averaging networks to perform a performance comparison with the intelligent chatbot, evaluating the performance of the three models on different tasks. Optimized the chatbot's capabilities for specific tasks under resource-constrained conditions. RELEVANT EXPERIENCE

A C++-based compilation evaluation platform, HUST C++, compiler Feb. 2024 - June. 2024 Ø Implemented a C language compiler. Developed core components of the compiler from scratch, including a lexical analyzer based on finite state machines and a syntax analyzer based on LR parsing. Ø Built a compilation evaluation platform for teaching based on the Educoder platform, which tests whether the code written by students correctly implements compiler functionalities, such as generating the correct three-address code or LLVM IR from the syntax tree, and whether code optimization is completed. Simulation Computer Implementation Based on FPGA, HUST C, Verilog Sept. 2022 - Dec. 2022 Ø Designed and implemented a computer system using the RISC-V instruction set with Verilog and Logisim, capable of supporting 29 basic instructions, multi-level interrupts, and simple exception handling. Throughout this process, core components such as a hardwired controller, a 5-stage pipeline CPU, and a Branch Target Buffer (BTB) and data paths were developed from scratch.

Ø Focused on optimizing the CPU by employing forwarding techniques and dynamic branch prediction, which reduced CPU performance loss due to data conflicts that caused pipeline stalls by 86%, with a branch prediction accuracy of 70%. Compared to the pre-optimization phase, the number of clock cycles required to run six benchmark programs decreased by 14% to 29%, and the CPU time was reduced by 11% to 23%. Ø Validated system scalability with an FPGA-based "flappy bird" game, reducing the hardware and software overhead, and interacting with the player via VGA display and mouse and keyboard. TECHNICAL SKILLS

Programming & Software: Python, C/C++, Java, pytorch, transformer, SQL/MySQL, HTML/CSS, Verilog, MATLAB, Logisim

Contact this candidate