I'm Jonas — a data scientist based in Melbourne. I build the models, pipelines and tools that take complex data and turn it into something a business can act on.
I enjoy working with numbers, finding patterns, looking at trends and visualise them in a way even people from non-technical background can appreciate. I like to ask why, why customers are satisfied, why the metric has changed in recent month, what are people talking about that's new. That takes me one step further — finding the specific driver behind a number, sizing how much it matters, and handing someone a tool they can keep using after I leave.
Day to day that means building and maintaining data-rich and interactive dashboarding, statistical modelling(regression, tree model, segmentation, factor analysis) for enterprise clients, integrating LLM into data pipeline that make unstructured verbatim data analysable at scale, and automation that quietly removes hundreds of manual hours for a client.
A benchmarking tool for a client that leverages LLM to classify hundreds of thousands of customer verbatims from multiple sources including survey and social review by journey stage and theme, then measures what each theme does to NPS. Built end to end — pipeline, model, interface, auth.
Open the walkthroughA multilingual dictionary project I build and maintain on my own time. Combines a dictionary and a Anki-style flash card system to become the one-stop platform for language learners. Incorporates full keyboard manoeuvrability inspired by vim-style navigation. Demonstrates my NLP, database management and LLM integration capability.
Visit dictdict.comQualtrics (formerly InMoment), Melbourne
InMoment, Melbourne
Octavius Trading, Hong Kong
Innovest Financial Group, Hong Kong
Client names and figures in the case study are anonymised. Built and hosted by me.