Yuchen Du
Sydney,NSW
Summary
Proactive AI engineer with strong academic training and hands-on experience designing, training and deploying large language models. Demonstrated success in improving Verilog code generation accuracy by 41 % through LoRA fine-tuning and optimising inference with vLLM and RAG pipelines. Adept at translating complex business requirements into scalable, cost-effective AI solutions using Python, Docker/Kubernetes and cloud platforms. Actively seeking to leverage these skills to deliver high-impact generative AI products for diverse clients.
Overview
3
3
years of professional experience
Work History
LLM Research Engineer (Capstone Project)
UNSW Sydney
02.2024 - 10.2024
- Engineered an end-to-end pipeline to fine-tune Llama-2/3 and CodeGen models using LoRA, boosting Verilog code-generation correctness from 51 % to 72 %.
- Integrated Retrieval-Augmented Generation and LangChain to supply domain-specific context, cutting hallucination rate by 30 %.
- Deployed optimised vLLM inference server on AWS GPU instances, achieving 4× lower latency and 38 % cost reduction versus baseline.
- Containerised training & inference workflows with Docker and Kubernetes; authored CI/CD scripts to automate model updates and batch processing jobs.
- Communicated findings in a 45-page technical report and presented to a panel of academics and industry mentors, receiving Distinction.
VR/Game Development Intern
Chesstar Information Technology
01.2022 - 04.2022
- Developed gameplay logic and data-driven UI in Unity (C#), demonstrating adaptability to new frameworks and agile collaboration across four-member team.
- Automated regression testing scripts that cut quality-assurance cycle time by 15 %.
Education
Master of Information Technology - Artificial Intelligence
University of New South Wales (UNSW Sydney)
Sydney, NSW01.2026
Bachelor of Science - Computer Engineering
University of New South Wales
Sydney, NSW11-2024
Skills
- Programming: Python, Java, C, SQL
- AI & ML: PyTorch, Hugging Face Transformers, vLLM, RAG, LangChain
- MLOps: Docker, Kubernetes, AWS (EC2, S3), Git, CI/CD
- Data: MySQL, Pandas, NumPy; Data processing and analysis
- Communication: Technical writing, stakeholder engagement, Agile teamwork
Accomplishments
- Kaggle Competition: Titanic – Top 20% (2024)
Participated in the "Titanic: Machine Learning from Disaster" competition on Kaggle, achieving a position within the top 20% by developing a predictive model using logistic regression and decision trees. - AWS Academy Cloud Foundations Certification (2024)
Completed the AWS Academy Cloud Foundations course, gaining foundational knowledge in cloud computing concepts and AWS services.
Timeline
LLM Research Engineer (Capstone Project)
UNSW Sydney
02.2024 - 10.2024
VR/Game Development Intern
Chesstar Information Technology
01.2022 - 04.2022
Master of Information Technology - Artificial Intelligence
University of New South Wales (UNSW Sydney)
Bachelor of Science - Computer Engineering
University of New South Wales
Similar Profiles
George KudrnaGeorge Kudrna
Senior Research Fellow at UNSW SydneySenior Research Fellow at UNSW Sydney
Quinton Vea VeaQuinton Vea Vea
Indigenous Cultural & Intellectual Property Manage at UNSW SydneyIndigenous Cultural & Intellectual Property Manage at UNSW Sydney
Lena CalligerosLena Calligeros
Senior Recruitment Product Manager at Future Student Recruitment, UNSW SydneySenior Recruitment Product Manager at Future Student Recruitment, UNSW Sydney
Thomas RobbThomas Robb
Qualified Paramedic at NSW Ambulance ServiceQualified Paramedic at NSW Ambulance Service
Yunwen MAYunwen MA
Account Manager at TradeMax GlobalAccount Manager at TradeMax Global