Building HeatSync Part 1: The Origin Story

Like most swim parents, we print heat sheets before each meet and use highlighters to mark our child’s events. The problem is that on race day, there are always so many kid-related things to think about and take care of, so the printed heat sheets often get forgotten at home—not remembered until right before the meet starts😩

In a recent meet, I forgot my heat sheet yet again. I had some work to do and wanted to use the gaps between events to work, but was afraid of missing my child’s races. In a pinch, I threw the heat sheet PDF file at ChatGPT and asked it to remind me 5 minutes before each of my child’s events.

ChatGPT’s approach pleasantly surprised me: it first found all my child’s race times from the PDF, then directly created reminder events in my calendar. The entire process didn’t require any third-party tools.

I shared this method with other parents in the club, and everyone found it quite practical. But many haven’t had much exposure to AI and lacked confidence in whether they could successfully operate it—even though the prompt I provided was already quite simple. And honestly, having to manually go through the process each time was indeed quite troublesome.

As a developer, this felt like the perfect practice project. A real problem I personally have, with users I can easily reach (my kid’s swim team), and a clear path to a solution.

So I decided to build it.

Why This Problem?

Swim meets are chaotic. Parents are juggling:

• Multiple kids in different age groups
• Events scattered across hours (sometimes days)
• Warm-up times, check-in times, actual race times
• Other kids to watch, snacks to manage, work emails to answer

The heat sheet PDF is the source of truth, but it’s designed for officials, not parents. Dense tables, tiny fonts, hundreds of entries. Finding your kid means scanning through pages, and even then you might miss an event because “100 Free” looks a lot like “100 Fly” when you’re squinting at a printout in a humid pool deck.

The current solution (print, highlight, carry) has a single point of failure: the human.

The Insight

When I watched ChatGPT work through the PDF, I realized something: this is exactly what AI is good at. Structured data extraction from messy documents. The heat sheet isn’t unstructured—it follows a format. Event numbers, heat numbers, lane assignments, names, times. It’s just tedious for humans to parse.

If AI can do this once through ChatGPT, it can do it a thousand times through an API.

The Decision

I could have stopped at “use ChatGPT manually.” But that’s not a product—it’s a workaround. I wanted something my non-technical swim parent friends could use without thinking. Upload the PDF, type your kid’s name, get calendar events. Done.

That’s HeatSync.

In the next post, I’ll walk through how I defined the MVP scope and set measurable goals before writing any code.

This is Part 1 of a series on building HeatSync. Part 2: Defining the MVP →