About Me

Personal Image

Here is Chenyuan Yang (Ryann, 杨晨源).

I am currently a 4th year student at the School of Computer Science at Fudan University, China. I’m going to pursue my master’s degree in the Language Technologies Institute at CMU.

My research focuses on Adaptive Scaling for Large Language Models (LLMs). Specifically, I am interested in developing architectural and algorithmic interventions — such as Mixture-of-Experts (MoE) routing and reinforcement learning-based sampling — to optimize how models dynamically allocate computation for complex reasoning and multilingual tasks. My goal is to build efficient, domain-agnostic systems that maintain robustness across diverse modalities, including mathematical reasoning and time-series forecasting.

Looking ahead, I aim to join a leading technology company as an NLP engineer, not merely implementing existing frameworks, but pioneering novel applications that directly address critical product challenges and enhance user interaction efficiency. In the long term, I aspire to become a machine learning tech lead, designing innovative technology solutions that bridge the gap between academic research and impactful product deployment.

I am actively seeking Summer Research for 2026 summer and beyond. If you have any information, please contact me. Thank you!

EDUCATION

Carnegie Mellon University, Pittsburgh, USA

Master of Science in Artificial Intelligence and Innovation

Incoming Graduate Student (Fall 2026 Entry)

Fudan University, Shanghai, China

Major: Computer Science and Technology | Sep 2022 – Jun 2026

Cumulative GPA: 94/100
Honors and Awards: Fudan University Outstanding Student Scholarship: Second Prize (2024–2025); Third Prize (2023–2024, 2022–2023)
My Transcript: transcript

University of California, Davis, CA, USA

Exchange Program | Sep 2024 – Dec 2024

Cumulative GPA: 4.0/4.0
My Transcript: transcript

PROJECTS AND RESEARCH EXPERIENCE

Fudan University Natural Language Processing Group

Research Intern
Shanghai, China | Jun 2024 – Present

Engineered NLP data pipelines by implementing text normalization, entity recognition, and semantic filtering techniques, to support model training for sentiment analysis and text summarization tasks.
Architected a safety alignment module using PyTorch, implementing content filtering (keyword blacklists) and rule-guided reward scoring based on safety criteria to enhance the safe text generation capabilities of language models.
Established a multidimensional evaluation framework combining automated rule-based scoring with human supervision to assess model outputs across safety, relevance, and instruction-following capabilities.

THINKDRIVE Autonomous Driving and Digital Twin System

University Student Innovation & Entrepreneurship Development Program
Web Front-end Developer
Shanghai, China | Sep 2024 – Present

Built core modules of digital twin web interface using modern JavaScript frameworks, implementing real-time vehicle status visualization and interactive controls.
Optimized front-end architecture with modular design, enhancing code maintainability across 20+ reusable UI elements; developed data bridging solutions between browser clients and autonomous driving simulators using WebSocket APIs.
Contributed to simulation platform development by creating configurable training scenario templates.
Adopted as a teaching tool in Principles and Practice of Artificial Intelligence for Autonomous Driving, an AI course at Fudan University.

Shanghai Library Open Data Competition

Team Leader, Web Front-end Developer
Shanghai, China | Jun 2024 – Aug 2024

Developed award-winning cultural heritage platform with 3D building visualization and AI-powered historical Q&A system.
Implemented responsive frontend using CSS grid layouts and dynamic data rendering (15+ interactive timeline modules).
Built RAG pipeline integrating Shanghai Open Data APIs and custom vector database, improving semantic search relevance.
Won the Excellence & Popularity Award.

INTERNSHIP

Bytedance
Data Governance Intern
Shanghai, China | Sept 2025 – Present

Developed a Unified Data Taxonomy by analyzing TikTok backend assets through exploratory data analysis, feature extraction, and sample validation, and identified patterns and correlations to guide standardized recognition rules.
Integrated algorithmic and LLM-based methods to optimize classification accuracy and recall, increasing classification capability coverage from 87% to 94% and improving classification performance (F1 score) from 92% to 95%, while supporting model enhancement and deployment through data pipeline construction.
Defined asset selection rules and implemented an SDLC-based incremental labeling process, addressing cold-start challenges for new business domains through LLM-based reasoning and few-shot learning, enabling sustainable and lightweight tag updates.
Designed the supporting analysis framework and applied business research and sample analysis to provide accurate references for automated recognition.

Transwarp Technology (Shanghai) Co., Ltd
Software Testing & Development Intern
Shanghai, China | Dec 2024 – Mar 2025
Executed functional and performance testing for ArgoDB distributed database across Linux/Windows environments.

Automated test case execution using Python scripts, covering 20+ core features including index management.
Analyzed system performance under concurrent operations, identifying bottlenecks and providing actionable feedback to optimize database throughput and latency.
Documented reproducible defect scenarios with JIRA issue tracking to resolve issues affecting multi-node consistency and fault tolerance.
Assisted in test environment setup and CI/CD integration using Git and Docker to improve test efficiency and deployment reliability.

EXPERIENCE

Class President, Computer Science Undergraduates Class of 2022, Fudan University
Shanghai, China | Sep 2023 – Present

Assisted the counselor in managing a class of 120 students, coordinating events such as spring outings, choir competitions, and class retrospective video production.
Streamlined class operations, strengthening teacher-student connections, and fostering a cohesive environment.

Math Tutor - Middle School Level ｜ Jul.2024 - Present

Provided one-on-one mathematics tutoring over 1.5 years, previewing new topics and designing targeted exercises.
Guided the student to tackle mathematical problems using visual modeling, structured decomposition, and step-by-step reasoning frameworks, which led to sustained improvement in mathematical performance and transferable analytical skills.

FISS (Fudan University Summer International Program) Buddy ｜ Jul.2024 - Aug.2024

Assisted exchange students by coordinating logistics and providing cultural guidance for a welcoming environment.

Ceramic Art Association “Sunshine Home” Practice Activity ｜ Mar.2023

Guided individuals with intellectual disabilities in ceramic-making, fostering creativity, patience, and inclusivity.