top of page
ReactCon_icon.png
ReactCon_logo.png

Patrick-mary Nwakeze

Senior Software Engineer (React Native)
Unidosh
greg-fawson-foto.1024x1024 Kopie.jpg

I found my way into software engineering through curiosity about the products people actually use every day, and that curiosity hasn't stopped. After helping build Finnesser, a fintech startup where I developed core financial features using React, Next.js, and backend APIs, I deepened my focus on frontend systems that scale. Today I specialise in React, Next.js, and React Native, using TypeScript and AWS to build maintainable, production-grade applications. I'm particularly drawn to the intersection of AI and data visualisation, building tools that help users make smarter decisions without overwhelming them. I've spoken at internal engineering sessions and local meetups, and I'm now bringing those conversations to larger stages. My talks are practical and story-driven: I start from real problems, show the messy middle, and leave audiences with patterns they can use the next day.

Replacing a Filter UI with Natural Language: What We Learned Shipping It

Session

Natural language search sounds better than dropdowns. Users say so. Product managers love the demo. Then you ship it and discover that "better" is doing a lot of work.At Finnesser, we replaced a multi-parameter financial filter UI with an LLM-powered query system in our React Native app. The demo was impressive. The first month of production data was humbling. Users didn't trust AI-generated results. Queries returned plausible-but-wrong answers. Nobody knew what the system could understand.This talk covers the full journey: the initial React Native architecture for streaming LLM responses, the three UX failures production exposed, the second-iteration patterns that stabilised it — confidence indicators, result explanation UI, structured filter fallbacks — and the honest verdict on when natural language UI is a genuine improvement versus a feature built for the demo reel.

bottom of page