Hello, I’m Bryan. I’m a lead UX designer at Eptura, a SaaS company that makes workplace management software. For over nine years I’ve designed their web apps, mobile apps and room screens. I work remotely from London.

I also design for AI and build with it. I code my own working prototypes and tools.

Next, I’m looking for a lead or senior product design role where I can keep designing, prototyping and shipping.

Selected work

Back to all work

2026

Throughline iOS app : What did I actually do?

An iPhone app I designed and built with AI, having never written code. Log a work win as it happens, then turn the record into CV-ready language when you need it.

Team
Bryan Hayward Designer and builder Claude Code
Contribution
IdeationStrategyUX designDevelopment
Platforms
iOS
Status
In testing In daily use by me; not yet released.

In short

Throughline is a small app for a common problem: remembering what you did at work, months later, when someone asks. You log a win in under a minute, at the moment it happens. When review season or a job application comes round, the app turns that record into language you can use.

The app is half the story. I designed it, then built it with Claude Code as my programming partner, with no coding experience of my own. It took weeks, not months. [CHECK]

How it was: Writing your own history from memory

Most designers are good at doing the work and poor at keeping a record of it. Not because they don’t care. The moment passes, the next brief arrives, and last quarter’s wins are already hard to reconstruct.

Then appraisal season comes round, and you’re digging through chat threads and old files for evidence of things you lived through but can’t quite describe. The numbers get vague, the language goes soft, and good decisions go unmentioned because you can’t remember why you made them.

There are apps for writing CVs, tracking performance and taking notes. I couldn’t find one that captured a win in plain language as it happened and turned it into something useful later. So I made one.

The ask: A brief I wrote for myself

The design problem wasn’t the gap in the market. It was getting people to log anything at all.

Productivity apps tend to fail at the same point: the gap between meaning to do something and doing it. If logging takes more than a minute, it won’t happen often. If the app asks too many questions, people give up before saving anything. If the output isn’t usable straight away, the app becomes a diary nobody reads. Those are habits to design for, not features to add.

What I did: Design first, then build with AI

Capture comes first
The app opens on a large text box and three small fields: project, type of win and whether a senior person signed it off. The prompt does the framing: “Describe the win. Lead with the outcome.” That line primes people to write in a way the app can use later, without making them think about CVs in the moment. A rough entry that exists is worth more than a perfect one that doesn’t.

The log is evidence
Each entry shows its project, type and sign-off as tags, at the same weight as the date. They aren’t decoration. They’re what the app reads when it groups and weighs entries later, so the person doesn’t have to.

A grid that rewards turning up
The Cadence screen shows one square per week, filled when you log something. Seeing a run of filled squares is encouraging in a way an invisible habit isn’t. An empty grid in week one is deliberate, and the copy keeps it calm: “Steady. The record is taking shape.”

The payoff
The Synthesise screen turns the log into CV language. You choose whether it leads with achievements or skills, whether it groups by project or by skill, and which role you’re writing for. Claude, through the Anthropic API, writes a draft from your entries in the right tone for that role. You edit rather than start from a blank page.

Building it
I did the design work before any code: the problem, the features, how each screen behaves and every line of copy. That became the brief for Claude Code. I decided what to build, in what order and why. Claude Code wrote the code, and I tested it, questioned it and sent it back when it wasn’t right.

[CHECK: add one specific moment — something Claude Code got wrong, and how you spotted and fixed it. One real example will carry this section.]

AI didn’t make the design problem smaller. It made the distance between a design and a working app much shorter.

[CHECK: tech stack — live page says native Swift; notes say a web build with a Node.js backend and a planned Swift rebuild. State whichever is true today.]

Where it stands: In my pocket, not yet in a shop

[TEST block – see 00-testing-framework.md]

Method
My own daily use since [CHECK date]; [TEST if anyone else has tried it]

Entries logged
[FIGURE]

What I’ve changed because of it
[CHECK]

Throughline is in testing on my own phone and not yet released. That’s deliberate. The test of an app like this is whether the person who built it trusts it enough to rely on it, every week. [CHECK: next steps — keep or drop the job tracker, interview prep and gap analyser ideas]

In hindsight: Writing the brief yourself

This project started with a problem I had and nobody had solved. That’s the same instinct behind most of the work in this portfolio: noticing the problem before anyone hands it to me.

The difference this time is that I could build the answer too, though not alone.