Research Swipe

Inspiration

As a beginner trying to get into AI/ML research, I had no idea where to start. Who are the important researchers? Which papers should I read? And even when I found a paper, I didn't know how to actually learn from it — not just skim it. That frustration became the spark for Research Swipe.

What it does

Research Swipe lets you discover the top 101 AI researchers worldwide, explore their work, upload any of their research papers (PDF), and instantly get 10 AI-generated study questions to test your understanding. You can choose your preferred AI provider (OpenAI, Google Gemini, Anthropic Claude, or Groq) and pick from multiple models. It turns passive reading into active, structured learning.

How we built it

Built entirely on MeDo through natural language conversation and iteration. I described the features I wanted — the researcher grid, the swipe cards, the PDF upload flow, the study session modal, the AI provider selector — and MeDo generated and refined the full-stack application. No traditional coding setup required.

Challenges we ran into

The first challenge was adapting to a completely new platform (MeDo) while having previously built on 4 different tools. The bigger challenge was the core problem itself — figuring out how to meaningfully combine four distinct things into one coherent experience: a curated researcher directory, a paper-based quiz engine, an answer evaluation system, and a multi-provider/multi-model AI selector. Getting all of these to feel like one seamless app rather than four separate features was the hardest design and product problem to solve.

Accomplishments that we're proud of

A clean, functional app covering 101 real AI researchers with photos, affiliations, and research areas A working PDF-to-questions pipeline that generates genuinely insightful study questions Support for multiple AI providers and model selection in a single interface Shipping a full product on a brand new platform within the hackathon timeline

What we learned

That the best product ideas come from real frustrations. I learned how to think about combining multiple complex features into a single user flow, and how powerful MeDo is for turning ideas into working applications without getting blocked by technical setup.

What's next for Research Swipe

User accounts with XP tracking and progress history across papers Leaderboards to compete with others on mastering research Expanding beyond AI to cover researchers in biology, physics, and economics AI-powered researcher recommendations based on your study history

Built With

• medo