I’m a blind AI engineer. I build accessible AI products for a living — and use them to work, every day.

Lived experience as rigour, not a sympathy hook

Automated scanners catch a slice of accessibility problems. Sighted auditors working from a checklist catch more. But the failures that actually stop a blind or low-vision person — a control that VoiceOver skips, a live region that never announces, a voice flow that traps you — usually survive both. I find them because I hit them, with real assistive technology, the way your users will.

What’s rare is the second half: I’m also an AI engineer with 16+ years shipping production systems across government, finance and payments. So when the product is an AI product — a chatbot, a voice assistant, an agent UI — I can review both the accessibility and the way the AI behaves for someone using it by ear.

What I do

• Manual WCAG 2.1 / 2.2 AA audit with real assistive technology

  I test your site or app with VoiceOver and a keyboard and give you a written, prioritised findings list — each issue, why it matters, and how to fix it.

• AI-product accessibility review

  Does your chatbot, voice interface or agent UI actually work for a screen-reader or low-vision user? This is the gap most accessibility consultants can’t cover, and the one I know best.

• Embedded advisory during a build

  I join your team as the accessibility voice while you’re designing and shipping, so accessibility is built in from the start rather than bolted on at the end.

• Screen-reader and VoiceOver testing of specific flows

  Point me at a checkout, an onboarding, a form or a key journey and I’ll test it end to end on a real device and tell you exactly where it breaks.

• Team training

  Practical, plain-English sessions on accessible design and development for designers, developers and product managers — grounded in how people actually use assistive technology.

How we’d work together

Three simple ways in — pick whichever fits where you are:

• Fixed-price audit — a defined scope, a set price, and a written report you can act on.
• Day-rate or monthly retainer — ongoing accessibility review through a build or across a product.
• Book a 15-minute scoping call — no quote gate, no pressure; we work out whether I’m the right fit for what you need.

Not sure which one? Email me a sentence about your product and I’ll suggest the lightest engagement that would actually help.

The proof is a product I actually shipped

I don’t have a wall of client logos to show you, and I’m not going to invent one. What I have is ChatLemur — an accessible, voice-first AI assistant that’s live on the web today, designed screen-reader-first and tested with VoiceOver and full keyboard navigation as its baseline. You can try it in your browser and judge the accessibility for yourself.

There’s also a native iPhone app, currently in TestFlight beta and still in development — on-device machine learning for a live camera perception aid, built and used by someone who navigates by ear. The whole thing is beta, built in the open, and I say so plainly; honest about what’s finished is exactly the standard I’d hold your product to.

What I will and won’t claim

I sell expert manual review — not automated fixes. No tool, mine or anyone else’s, can make your product “compliant” by bolting on an AI overlay, and I won’t pretend otherwise. Operators who sold that promise have been fined by regulators and rejected by the blind community for good reason.

So, plainly: an audit is not a compliance guarantee and not legal protection. What it gives you is an honest, expert picture of where your product stands against WCAG 2.1 / 2.2 AA and a clear, prioritised path to fix what matters most. The decisions, and the remediation work, stay yours — I’ll help you get them right.