Resume
Experience
-
Digitalization Manager Hong Kong Productivity Council
2024 - PresentWork as an AI expert and solution architect in streamlining various parties to initiate and landing innovation and digitalization projects in HKPC. The followings are summaries for the works:
- take lead in a GenAI project utilizing SOTA RAG techs to create a solution to transform complex funding scheme vetting procedures to bring substantial productivity gains.
- lead and oversee cycle time optimization project. -
Senior Data Scientist Apoidea
2022 - 2024Work as a small team lead of data science. Take a lead role in development of quant trading project and be Responsible in leading development of RecSys and engage in development of LLMs empowered information aggregator platform. The following are some signature achievements or contributions :
- utilize SOTA and efficient deep RL algorithm to develop a intelligent quant system reaching no-man or near no-man condition by continuous learning and self-adjustment capability of the system.
- The quantitative strategies includes statistical arbitrage and trend-following, which all reach more than +50% annualized return and Sharpe Ratio more than 3.
- The leading role in RecSys include designing, implementation and orchestrating of a production grade personalized recommendation system. The resultant Recsys reach 0.4 in nDCG@10.
- Engage in strategic planning of the development road path for RecSys and the platform as a whole.
- Lead and conduct A/B testing pipeline and statistical test on recommendation system evaluation.
- Engage in News aggregator platform development powered by latest LLM and NLP techs, including prompt engineering pipeline, model development. -
Data Science Lead Wonderbricks
2021 - 2022- Lead a team of data scientist to build a deep recommendation engine for material substitution and a generative design engine for building.
- Transform commercial concepts, needs and requirements to suitable and robust ML solutions.
- Initiate and create master plan for ML technology development alongside with the business development need of the company.
- Responsible for ML solution and system design from 10k to 100k users scale.
- Design MLOps framework and pipeline for ML services.
- Lead R & D projects for ML service development with domain experts.
- Build up a effective task management system for a team management. -
Data Scientist Gammon Construction Limited
2019 - 2021- Lead and be hands-on for the development of data science projects in construction, mentor junior members and be responsible for AI solution architecture design.
- Manage AI project productization and bridge the communication between users and technical team.
- Build up data science project development practices and processes.
- Lead the team to create a competitive computer vision platform for 7 tasks specific for construction business equipped with various cutting-edge deep learning models in computer vision. The platform is scale-up ready, flexible, adaptable to edge-device and production ready.
- Build a transformer tech empowered NLP system for sequence classification to automatic categorize safety issues in construction, demonstrating productivity gain and reduction of fatal cases. The resulting model improve more than 70% over the original one.
- Lead a POC project for creating an intelligent scheduling system for concrete production and delivering by utilizing deep reinforcement learning for it's self-learning capability and adaptability for complex conditions.
- Build an unsupervised deep forecasting model to predict concrete demand, which embedded into a web application. It is demonstrated to be able to reduce 40% of the order error, resulting in a large amount of cost reduction.
- Lead the team to develop a generic library for continuous development of machine learning technology. -
Data ScientistRobot Data
2018 - 2019- Lead R&D projects in computer vision for AI-robotic solutions.
- Engage in pitching with client and investors for AI
- Be responsible for solution architecture design and end-to-end hands on implementation for computer vision.
- Build a competitive facial recognition system and an airplane repairing parts recognition system. -
Machine Learning EngineerKami
2018 - 2018- Engage in the development of a intelligent conversational
- Be responsible for the R&D of knowledge graph embedding model for link prediction and graph completion.
- Engage in development of generic NLP and machine learning library.
Education
-
The Chinese University of Hong Kong
2012 - 2017Bachelor of Science in Mathematics (HON)
My skills
-
Python
-
Reinforcement Learning
-
Recommendation System
-
NLP, LLMs Finetuning, Prompt Engineering
-
Generative AI, ChatGPT
-
Quantitative Trading, Statistical Modeling
-
Data Engineering, MLOps
-
Computer Vision
-
MongoDB
-
Amazon AWS