The situation
A growing health-tech startup wanted to use AI and better analytics. Its data had grown with the product: useful, but spread out, inconsistent, and hard for product and business teams to reach. Before building anything clever, the company needed to know what it actually had.
What I did
- Advised on AI adoption. I worked with the team on data analytics, machine learning and data architecture, so that AI decisions were based on what the data could really support.
- Audited the database. I led a full audit of the PostgreSQL database: where the data comes from, where the quality gaps are, and who uses what.
- Redesigned the data architecture. Together with the team, I designed a new, scalable architecture for advanced analytics, reporting and easier data access for product and business teams.
- Built reporting and research. I produced monthly reports and wrote a wider research report on the company’s medical data.
- Delivered LLM-based solutions on top of the cleaned-up foundations.
What changed
The team now has a clear map of its data and a plan for where it is going. Product and business people can get to the numbers they need, and AI work starts from foundations the team understands and trusts.
Tech used
PostgreSQL, Python and SQL, data modelling and architecture, LLMs, and regular reporting.