I am a seasoned Data Scientist and Machine Learning Engineer with over 10 years of experience in end-to-end ML system design, scalable deployment, and MLOps. I hold a PhD in Machine Learning & Artificial Intelligence and have published in top-tier conferences (e.g. ICML, NIPS, IJCAI, ICDM etc.). My expertise spans classical & deep learning, generative AI, Azure AI/ML, and production-grade APIs.
I have worked on a wide range of projects—from building digital twins and predictive models to integrating large language models into real-world applications. I enjoy turning complex problems into practical, scalable solutions, and has collaborated with teams across industries like water treatment, 3D point cloud, material discovery, health data analysis and cybersecurity. I am passionate about making AI useful in everyday business and brings a thoughtful, hands-on approach to his work.
- 📍 Melbourne, VIC, Australia
- 📧 lichengxlxl@gmail.com | 📞 +61 420 714 881
- 🎓 PhD (Deakin University, Australia)
- 🏅 Azure AI Engineer Certificate
| Repository | Summary | Tech Stack | Link |
|---|---|---|---|
| MCP-Autonomous-Agent-System | An end-to-end autonomous agent system powered by the Model Context Protocol (MCP) | Python, FastAPI, Node.js, MCP, LLMs | Repo |
| OZlaw AI | AI-driven legal document analysis and search tool with memory. | Python, Django, LangChain, OpenAI API, ChromaDB | Repo |
| Intelligent-MultiAgent-Document-and-Web-Search | Multi-agent document and web search system using AI agents. | Python, FastAPI, LangChain, LangGraph, OpenAI | Repo |
| prompt_comparsion_system | Framework for LLM prompt comparison and evaluation. | Python, FastAPI, LangChain, Hugging Face, MLflow | Repo |
| OULAD_prediction | Weekly student performance forecasting on clickstream data. | Python, LSTM | Repo |
| IMV-LSTM-Classification | Interpretable multi-variable LSTM for time-series analysis. | Python, PyTorch, IMV-LSTM | Repo |
| highdimBO_dropout | High-dimensional Bayesian optimization with dropout. | Python, GPflow, Bayesian Optimization | Repo |
Senior Data Scientist, Trility (03/2022–02/2025)
- Architected a Retrieval-Augmented-Generation chatbot using Llama Index, LangChain, Hugging Face, MLflow & Azure ML (93% hit rate, low latency).
- Built deep learning chemodose time-series predictors with MLflow & Azure ML (+8% accuracy).
- Developed a Julia-based digital twin for wastewater simulation (100× speedup, AUD $30k savings).
Senior Data Scientist, Euclideon (05/2021–03/2022)
- Led a 3D point-cloud segmentation platform in C++/PyTorch on Azure ML (+15% accuracy).
AI Researcher / Data Scientist, NUS (08/2019–03/2021)
- Implemented VAE/GAN models for network anomaly detection (+20% accuracy).
- Designed predictive models for student performance (+12% success rate).
Machine Learning Researcher, Deakin University (08/2015–08/2019)
- Invented high-dimensional Bayesian optimization methods for polymer synthesis (–30% experiment time).
- Advanced toxicity prediction via multi-instance & multi-task learning (+4% over baseline).
- Deep Learning: PyTorch, TensorFlow, LSTM, CNN, VAE, GAN, Point-Cloud Segmentation
- Machine Learning: Bayesian Optimization, Time-Series Forecasting, Clustering, AutoML
- MLOps & Deployment: CI/CD, MLflow, Streamlit, FastAPI, Azure AI/ML, Docker, Git
- Generative AI & LLM: Transformers, LangChain, OpenAI, Hugging Face, Agentic AI
- Programming & Tools: Python, Julia, C++, SQL, MATLAB
I’m always open to collaboration, consulting, or new opportunities. Reach out via email or LinkedIn!
README generated by combining professional profile with GitHub portfolio.