About Me
Hello! I’m Chenchen Feng, a second-year master’s student in Computer Science at the Southern University of Science and Technology (SUSTech). My work focuses on Evolutionary Computation, backend engineering, and LLM applications. Since February 2026, I have been interning with the Harness team at Moonshot AI (Kimi).
- Location: Shenzhen, China
- Institution: Southern University of Science and Technology (SUSTech)
- Email: chenchenfengcn@gmail.com
Education
Master of Computer Science, Southern University of Science and Technology (SUSTech), 2024-Present
Bachelor of Computer Science, Southern University of Science and Technology (SUSTech), 2020-2024
- Thesis: “MetaDE: Evolving Differential Evolution by Differential Evolution”
Work Experience
- Harness Development Intern, Moonshot AI (Kimi)
February 2026-Present- Develop agents for LLM products and turn model capabilities into reliable user-facing workflows.
- Focus on agent orchestration, tool-use flows, and engineering support for agentic applications.
- LLM Post-Training Intern, Sangfor Technologies Inc.
December 2025-February 2026- Worked on post-training for a security-focused large language model.
- Contributed to supervised fine-tuning (SFT) and reinforcement learning from human feedback (RLHF) workflows.
- Backend Development Intern, Kevin Software Information Services Co., Ltd.
October 2022 - January 2023- Designed and built an extensible hotel database using MySQL to facilitate development.
- Developed back-end features in C#, including user registration, room search, room booking, payment transactions, and order reviews.
Projects
- AI Academic Research Assistant (Python, RAG, Agents, Streamlit)
- Built a RAG engine with traceable citations to reduce hallucinations and designed a parallel Agent architecture to optimize retrieval speed.
- Developed a robust Streamlit Web interface with multi-layer degradation mechanisms, ensuring high availability for batch document processing.
- Supercritical Airfoil Design and Optimization
(Python, Matlab)- Designed and optimized a supercritical airfoil for a large wide-body passenger aircraft using deep learning.
- Achieved superior performance in flutter coefficient and lift-to-drag ratio compared to manual designs.
- Research Fund Management System
(Spring Boot, MySQL, Vue)- Developed a system for university faculty to manage fund applications and approvals, featuring data visualization, messaging, and high concurrency.
- P2P File Transfer with UDP and Congestion Control
(Python)- Built a UDP-based P2P file transfer system with TCP-like reliability using congestion control techniques.
Publications
- MetaDE: Evolving Differential Evolution by Differential Evolution - IEEE Transactions on Evolutionary Computation, 2025
- MetaDE is a GPU-accelerated evolutionary framework that optimizes Differential Evolution strategies through meta-level evolution. Code
- Unleashing the Potential of Differential Evolution through Individual-Level Strategy Diversity - IEEE Transactions on Evolutionary Computation, 2026
- iStratDE is a GPU-accelerated Differential Evolution framework that improves performance through individual-level strategy diversity. Code