Gnosis AI

The problem

Financial data is dense, fragmented and written for specialists. Understanding what a signal actually means — and what to do about it — takes expertise most people do not have.

What I built

An AI-powered financial intelligence platform that turns market and transaction data into plain, actionable output: BUY / SELL / HOLD recommendations, portfolio insight, anomaly detection and interactive visualisations, delivered through an interface designed for a non-specialist.

Prototype first. Work out what a non-specialist actually needs to see, decide which features earn their place and where they belong, build that version alone, then bring the team in to integrate the technical implementation behind it.

Team Lead · 4-person team

• Led and coordinated a four-person team through the full build.
• Defined how the product should work from a layperson perspective — which features were essential and where they belonged.
• Independently created the initial prototype before it was handed to the team.
• Designed and developed the frontend and UI, translating user needs into product workflows.
• Refined and polished the prototype, then coordinated its integration with the technical implementation.
• Reported progress and technical decisions to the project supervisor throughout.

Backend services and the machine-learning pipeline were a team effort. This was product leadership and interface ownership, not a solo build.

• Next.js
• Python
• FastAPI
• Scikit-Learn
• XGBoost
• SQL
• Vercel
• Render

Where it got to

Delivered as a working, supervised major project — a concept translated into a usable, polished product with a defined feature set and a coherent user experience.

The three questions that matter