At VentureSEA Digital Singapore, I built a chat agent with three kinds of memory, combining in-prompt notes, a searchable message log, and a notebook for long-form notes. I wired it together with Google Agent Development Kit and deployed a demo on Hugging Face Spaces.
At Australian National University, I created a multilingual Named Entity Recognition pipeline that processed more than 7,900 news articles and extracted over 8,500 entities. I also developed a LangChain evaluation framework and a quantitative quality gate that raised inter-model agreement from 0.789 to 0.897.
At Adidas ODM, I developed a cost analysis reporting system covering around 50 products and led a migration of historical product data to SAP. Earlier, I led production trials and transfers to mass production, applying PDCA and root-cause analysis to reduce cycle time by four times.
My projects include a Python and LangChain pipeline for consolidating multilingual organization names and a PyMC hierarchical Bayesian mixture model for consumer segmentation. I bring a manufacturing operations background and a Master of Applied Data Analytics from The Australian National University to building and evaluating practical AI solutions.
