A Week Inside Gauntlet: How I Built My CollabCanvas

Montessori mom turned Web3 enthusiast, sharing insights on DevRel, coding, and continuous learning. Passionate about tech, community, and personal growth.
Search for a command to run...

Montessori mom turned Web3 enthusiast, sharing insights on DevRel, coding, and continuous learning. Passionate about tech, community, and personal growth.
No comments yet. Be the first to comment.
Ever felt the pain of Web3 whiplash? You mint an NFT on one chain, earn tokens on another, and two clicks later you’re stuck — staring at a bridge fee that costs more than your loot. It’s like buying a wristband for every single ride in a theme park,...

There’s something oddly satisfying about combining two things I love—building with code and making everyday life feel a bit more intentional. Over a whirlwind week, I built ChainCal, a tool that turns smart‑contract events into calendar reminders, em...

If you’ve ever spent hours studying something only to forget it the next day… you’re not alone. And no, your brain isn’t broken. It’s just wired for something different: curiosity. As someone who spent over a decade working closely with how the brain...

As you might know by now, one of my favorite parts about working in Web3 and attending events like EthDenver is getting to meet incredible people who inspire curiosity. Recently, I had the pleasure of chatting with Pauline, the DevRel at Wormhole, an...

So, I joined Gauntlet AI, which basically means going from zero to a hundred in a few weeks to become the best context engineering possible. We’re talking about learning things that aren’t even out yet and spending a hundred hours coding. The first week’s challenge? Build a Collaborative Canvas MVP in 24 hours, and then integrate AI features by the end of the week.
Now, what made this interesting is that we weren’t just coding. We were context engineering. That means before I wrote any code, I used AI to run an ideation session. The AI asked me questions, poked holes in my ideas, and helped me create a PRD and a Mermaid diagram. This planning phase was crucial because it set the whole foundation so I wouldn’t just fall into “vibe coding” and clicking yes to everything.
For the MVP, we started simple: making sure two users could see each other’s changes in real-time without lag. Once that was done, we moved on to the AI integration. Initially, I tried using DALL-E 3 to transform images into Japanese anime style, but it didn’t quite work. The AI was just generating random people. Eventually, I switched to Replicate, which gave me a much better image-to-image transformation and finally got the anime effect we wanted.
In the end, the app became this cool tool where users can create e-invites, customize them live with AI-generated art, and share them. My biggest lesson? You only truly learn by doing. You can hear about PRDs and context engineering, but you don’t really get it until you jump in and start solving problems. So, if you’re building something similar, focus on that strong foundation first, plan well, and let the coding flow from there.