Masaki Kawakami
Data & AI · Tokyo ⇄ Sydney
Based in Sydney, working across Japan and Australia.
The work that is live or on its way, in Japan and Australia.
Scroll or drag
Kept separate from the client work: research and university projects.
AI agents can be hijacked by malicious instructions hidden inside the content they read, an attack called prompt injection. Warden watches the model's reasoning while it works and blocks the reply the moment the reasoning starts drifting toward an attacker's goal. Across 2,880 evaluation runs it cut successful attacks by more than a third while traditional defences did no better than nothing, with a third of their false alarms.
View project→Out of 14,774 player seasons, fewer than 1 percent get drafted, so the real job is ranking every player by how likely they are to be picked and finding the needles. My final model ranked them near-perfectly (0.9990 on the Kaggle public leaderboard), and the reusable pieces became my own pip package.
View project→Will it rain in Sydney seven days from now, and how much over the next three days? I trained a model for each question, then did the part most coursework skips: packaged them with Docker and put them on the internet as a working API. Modest accuracy, honestly reported; what this project shows is the road from notebook to running service.
View project→A decision-support tool, not a trading bot: it forecasts tomorrow's Bitcoin high so a trader can sanity-check a plan. The first model failed the way naive price models fail, and reframing the target (predict the change, not the price) cut test error by six times. A Dockerised FastAPI serves the model on live prices.
View project→Only 56 students in roughly 1,000 reach the top grade, and the university wants to spot them before final results are out. After forensic data cleaning and honest feature work, the tuned random forest found 10 of the 11 top students in the test set, with a single false alarm.
View project→Each of four team members owned one model family over a telco dataset; mine was anomaly detection, plus the usage-data cleaning everyone built on. Across 7,043 customers the flagged group was not single outliers but combinations of extremes, with ten times the refunds and thirteen times the extra data charges, reading as refund abuse, referral fraud or missed upsells.
View project→Raw files from ten countries land in Azure cloud storage, Snowflake reads them in place, and SQL turns them into one clean table of 2.6 million rows. The analysis ends in an actual recommendation: comedy earns the highest engagement of any category, and it works across countries.
View project→Eleven years of New York taxi trips, 964 million rows after cleaning, processed with Spark on Databricks. The analysis answers operator questions (where the money is, when tips happen, which trips pay a driver best), and a fare model closes the loop, cutting prediction error by 28 percent against the baseline.
View project→The same setup a company data team would run: raw files land in Google Cloud, Airflow triggers the work on schedule, and dbt transforms everything step by step into clean, query-ready tables. History is preserved so last month's numbers never silently change, and the whole pipeline reruns every month without breaking.
View project→The photos are real ones taken by blind users, so they are often blurry, dark or off-centre, and describing them is genuinely hard. I built two generations of the model end to end, measured the jump between them, and then looked inside the newer one to see where it looks in the image when it chooses each word.
View project→Out in the community.
The business side and the build side. I work across both.
I use AI to change how businesses run, and build the workflows behind it myself. As COO of Cubic Innov8, a cross-border innovation hub connecting Japan and Australia, I bring in the clients and run operations on the business side, then build the systems that serve them. As an independent venture, I designed, built and shipped Vacanti AI, an AI job matching SaaS, solo, all the way to production. Before this I spent five years in HR at Canon Marketing Japan, then completed a Master of Data Science at the University of Technology Sydney, and I am based in Sydney, working across Japan and Australia.
- The business sideBring it in, run itAs COO, I bring in the clients and keep the operations running.
- The build sideShipped to productionI design and ship the systems behind that work, solo, all the way to production.
- The bridgeHR, data, two marketsFive years of HR domain knowledge, a data science degree, and two markets: Japan and Australia.
From the first conversation to a system running in production. Carried by one person.
- ScopeScopeStart on the business side: find the paying client, then map where AI actually changes the work.
- DesignDesignShape the scoring, the SOPs and the workflows, working back from what the business needs.
- ShipShipTake it past the notebook, all the way to a production system real users depend on.
- OperateOperateLaunch is not the finish line. Measure, fix, and keep it healthy in production.
