I'm a developer and PhD candidate at The University of Auckland researching in "Natural Language Processing in Clinicial Trials". I working across a variety of fields to solve complex problems with technology. While my main focus is on AI and machine learning, I also take on projects in data science, software development, web development, and blockchain, depending on what interests me at the time or what the job requires.
I enjoy exploring different areas of technology, applying AI to create innovative solutions and combining it with practical skills in development to bring ideas to life. Whether it's building predictive models, diving into data analysis, or creating user-friendly applications.
Build robust and scalable applications, handling both front-end and back-end development for web applications, as well as developing software tools and systems. I’ve previously developed lending systems, integrated chatbots, worked on blockchain tools, and created interactive data dashboards to assist clients in gaining insights from complex data.
Offer services ranging from data analysis to data mining, data manipulation and webscraping. I help clients extract insights from large datasets, employing statistical methods and algorithms. My expertise also includes machine learning, where I implement predictive models that allow clients to perform classification and forecast future trends. I further extend my work into AI, creating chatbots, AI agents, and even image detection systems.
Design and deliver introductory courses on AI and Python programming in Hong Kong, specifically targeted at primary and secondary school students. I also coordinate with fellow instructors to continuously improve the course materials and teaching methods, ensuring the best learning experience for students.
Reference up on request.
Development of a high-throughput indexer to collect real-time data from Solana and Ethereum, and aggregate it for specific use cases. Lead the design and implementation of a message-broker system (e.g., RabbitMQ) to efficiently handle user operations and maintain system performance. Contribute to full-stack web development and oversee Docker containerization for the entire infrastructure. Manage the design of databases for performance and efficiency, ensuring smooth data flow and scalability. As the lead developer, play a key role in multiple aspects of the project, collaborating across teams to ensure the system's overall efficiency and reliability. Startup reached an approximate $5 million USD market cap at its peak.
Reference up on request.
By using clinical trials (CT), high quality decisions can be made to benefit the healthcare systems, but it can often become expensive and are prone to failure. With the large amount of clinical data from the past, more modern techniques can be applied to gain inference. Natural language processing (NLP) techniques are now widely used in different areas like grammar induction and text-to-speed transformation. However, NLP is not yet mature in the CT area. It would be beneficial to include NLP techniques to automate certain tasks like processing clinical reports to obtain important information and optimizing trial designs. This could potentially lower the cost and decrease failure rates substantially.
In the recent update of iNZight, a joining and reshaping module was introduced. This module provides a tool for users to join and reshape data sets. Though it is not difficult to apply these operations and obtain the result with a tool like iNZight, there is a lack of tutorials or tools that explain the underlying process of these operations in an easy to understand manner. In this report, why these data operations are useful and important will be discussed. Software will also be developed that attempts to teach key concepts associated with joining and reshaping data sets through animations.
Read more here.
Most current software for survey analysis reads the data into memory, however, most of these computations can actually be expressed as database operations. The aim of this package is to provide a set of functions which allows survey statistics to be computed directly inside a database. To do this, we used dplyr and dbplyr. Lastly, this package takes a step further, it also provides graphics computation directly inside a database, where data tables are read into memory only when necessary.
Read more here.