Vergel Evans

UX Architect and AI-native engineer, founder of Humanjava. 20+ years designing complex systems across banking, government, and developer platforms. Now building and shipping conversational-AI and open-source products with agentic AI, and writing about how people and AI actually work together.

UX Architect Ā· AI-Native Engineer Ā· Service Design Strategist


Vergel Evans is the founder of Humanjava Enterprises, a UX consultancy he started in 2004. Over the past twenty years he has worked across investment banking, government digital services, fintech, and open-source developer platforms, designing complex systems that people can actually use. His work spans the full spectrum: from enterprise platform transformations at CIBC and Aviso Wealth, to government service design for the Ontario Ministry of Transportation and the Government of Alberta, to designing and shipping AI-native software with agentic workflows.

He holds a certificate in Information Management from the University of Toronto (graduated with honours) and is an HCI-certified Usability Analyst through Human Factors International.


In my own words

I've been doing some version of this work since the early 2000s, starting with building websites for Toronto's techno music scene, which is how Humanjava got its name. Human for the people at the center of every design decision. Java for the technology that serves them. The name stuck even as the work got bigger.

Most of my career has been spent in regulated, high-stakes environments. Investment banking at CIBC, where I led strategic UX across desktop, mobile, and native platforms for three and a half years, then came back twice more for different lines of business. Government service design at MTO's Digital Garage, where we ran Google Venture sprints to rebuild paper-based systems into digital tools that field officers could actually use. Financial services at Aviso Wealth, where I took the Qtrade mobile app from a disjointed v1.0 to a consolidated v2.0 experience.

The thread through all of it: I'm the person who comes in when something is messy, complicated, or stuck, and I figure out how to make it make sense. Not just the screens, but the underlying workflows, the data models, the handoffs between teams, the places where the front-end experience and the back-end systems need to actually talk to each other.

What I’m building now

The last couple of years have been a shift. I still consult, I recently wrapped a contract with the Government of Alberta designing a dispute resolution platform for condo disputes, but I'm also building and shipping my own products, and writing the code.

VoxRelay is the clearest example of where I am now. It's a production conversational-AI voice platform that phones and texts a small business's clients to confirm, reschedule, and cancel appointments, in English and French. I built the real-time pipeline end to end: speech to text, a language model that runs the conversation turn by turn, and text to speech, fast enough to feel like a normal call. The agent is grounded every turn in the business's own data so it answers from real state instead of making things up, and a second model scores each call and catches confusion while it's still happening. It's live, and it takes real calls.

I'm also the platform and technical lead on the Native Communications Society of the NWT's language project, working through MonkeySphere AI. That's two pieces of work: I migrated their content-management system off an aging, security-poor stack onto a modern, better-supported one, and I'm building the platform for their Indigenous-language learning program. The community and its language experts own the language and cultural work; my part is the platform underneath it, built to be accessible from the data layer up.

NostrKey is my open-source calling card. It's a full suite for the Nostr protocol: a browser extension for Chrome, Firefox, and Safari on both macOS and iOS, plus mobile apps, built from shared source and shipped in five days using Claude Code and agentic AI workflows. I also maintain a 7-package TypeScript developer library for the Nostr ecosystem, published on npm. That's not hypothetical "AI-curious" work. It's production software with live users and real developer adoption, and anyone can go read the code.

Before that, I took Sound HSA from pitch deck to working MVP, a health savings account mobile app on open-source protocols and blockchain. I designed the full experience in Figma, evaluated banking service providers, and handed off to a dev team in Berlin. The MVP reached 1,800 registered users.

I placed 3rd at the NSALUD Decentralized Health App Challenge with WellnessAlly, and I've been writing about what I call LLMDX, Large Language Model Design Experience, which is my attempt to work out the design principles for how people and AI actually work together in conversational interfaces.

How I think about design

I think in systems. Not just user flows, but how products, platforms, APIs, data models, and business processes connect. When I'm working on a project, I'm always asking: what happens upstream and downstream from the thing we're designing? Who else touches this data? What does the developer need from us to build this? What does the business need to see to know it's working?

That's what makes me effective in platform work and complex enterprise environments. I'm not just designing screens, I'm designing the experience architecture that everything else hangs on, and increasingly I'm building it too.

The short version

20+ years of UX and service design, now running the engineering as well. Banking, government, fintech, open-source, and my own AI-native products. I build things, I ship things, and I'm always looking for the next hard problem.


LinkedIn: linkedin.com/in/vergel-evans
GitHub: github.com/vveerrgg
Writing: humanjava.com