BuddieFit, Mindscribe and TechPath: what I built, and the problems I solved on the way.

BuddieFit: building the foundation with the team

BuddieFit was a fitness app built around a problem the team wanted to address: staying consistent is harder without accountability, support and motivation. The concept was to help people find fitness partners with similar goals and interests.

I was one of two developers on a team of about nine. I helped research the problem space, set up the Bubble app’s database structure and components, and handled the API integrations. As we built, I also redesigned some screens and improved the product’s look and feel.

We built the landing and marketing pages first, then moved into sign-in and the other product pages. My co-developer was new to Bubble, so I taught them as we worked, which meant handing over some of the front-end build and supervising it while I focused on the foundation and integrations.

Testing before demo day

The team’s testing notes cover creating an account, setting up a profile, finding a buddy, messaging and sharing progress. They also record the issues we found, including a missing back-navigation flow and age-field validation that needed work. Both were fixed before demo day.

Over nine weeks, the team took the concept to a demo-day presentation.

Mindscribe: connecting writing support to the product

Mindscribe was an AI writing and journaling product. I joined the team later in its program and built the product in full, including its OpenAI integration.

The integration takes a user’s writing, sends it to OpenAI for improvement and shows the revised text in the app. I later documented the approach as a guide for other Bubble developers, covering the API Connector setup and the workflow that returns the result to the interface. The team presented Mindscribe at demo day.

TechPath: making a career change easier to navigate

TechPath was a career-transition MVP for people exploring a move into tech. It combined a career recommendation with courses, mentors, learning resources and communities related to the recommended path. I joined the product team later in its program and built the MVP in full in Bubble.

The recommendation flow took a résumé and a short personality test as inputs. The product used the information extracted from the résumé and the test responses to inform an OpenAI-powered career recommendation. Users could also pay through Stripe to access the full feature set.

The résumé problem

The AI integration we were using at the time couldn’t parse the content of PDF files, though it could read Word documents. Most résumés are PDFs, so this blocked the recommendation flow.

I spoke to traditional developers, who advised converting the uploaded file to base64 and sending it through OCR for analysis. That let the product extract résumé information from PDFs and pass it on to the recommendation step.

Three cohorts, three kinds of work

BuddieFit gave me time to build with a team from the start, set up the Bubble foundation and teach a teammate the tool. In Mindscribe, I joined later to build an AI writing product, and that work became a guide for other Bubble developers. In TechPath, I built a more involved integration connecting résumé data, OCR and OpenAI.

The outcomes I can verify are the products built and the teams’ demo-day presentations. These were accelerator projects, so I don’t claim launch, adoption or business results.