Sujin K

If one dev with Claude Code can ship a whole app in 3 months, what exactly are startups raising $10M for?

Hey I'm Sujin, building Vanta, a health OS as a React Native app. It pulls in your labs, wearables, supplements and habits, and an AI coach does the managing: turns it into a protocol and keeps you on it Let's connect if you're building mobile apps or with AI agents.

My calorie tracker needs the nutrition label of every packaged food. There's no clean source for it, so I'm scraping store websites. If your app runs on data you don't own, how do you get it? Scraping, an API, hand entry?

Claude Code verifies its own work badly. It reads two files and assumes the rest. So before every feature I run a probe agent first. It reads the actual code and docs and comes back with facts, then I write the plan from that. How do you stop it building on things it only thinks it checked?

Added 11,029 packaged foods to Vanta's calorie tracker today. Apparently every store website has a sitemap, a plain list of every page on the site, one per product (it's for SEO). I built a pipeline that pulled the product pages from the sitemap and had Haiku read the nutrition facts off the product images. Anyone else scraping for their app? Curious what's worked.

Is it just me or do animations always look janky on the iOS simulator and perfectly fine on the phone

Reached 540 followers today 🥳 Was stuck at 100 for two months, so this one feels earned. Building something? Drop it below and let's connect 👋

Beta users taught me something uncomfortable this week: the app I built was a feature list, not a product. Labs, wearables, supplements, protocols, habits, all in one place, and nobody knew where to start. So I'm cutting the UI down to calorie tracking and workouts. The model watches your weight trend and recovery and adjusts your macros and training as you go. Shipping it to the same beta users this week and watching engagement. Wish me luck.

Hey I'm Sujin, building Vanta, a health OS as a React Native app, with Claude Code and Cursor doing the heavylifting. First version is live with a small group of beta users. Learned fast that a generic health dashboard full of features is confusing, so I'm narrowing it down hard. Let's connect if you're shipping mobile apps or building with AI agents.

Just crossed 400 followers :') 300 yesterday morning, 421 tonight. Met more builders this week than in all of last year, and the conversations every day are the best part. Still can't believe how fast this is moving. Next stop 500. Let's connect if you're into: - AI agents / vibe coding - iOS / app development - SaaS - Founders building in public - Health, fitness, longevity

Builders: what's your actual test loop right now? a) unit tests b) click through the UI yourself c) AI writes the tests, you trust them d) users are the tests

The way I build with AI keeps changing. Biggest shift so far: the order. How I used to build: build the backend → build the frontend → test → something breaks → change the backend → change the frontend → repeat How I build now: plan slices (frontend + backend together) → build one slice → test it in the frontend → fix → next slice AI can write the code and review the code. You'll only ever know if it's working when you test it through the UI. How do you go about ordering it?

60 new followers in one day :') Woke up at 300, at 360 now. This community shows up for each other, and that's the best part of being here. Solo founder, building Vanta, a health OS, in the open. Looking to connect with more devs and founders! Let's go!

honest question for people building with claude: do you read every diff, or just run the tests and move on?

Spent the last 2 months building Vanta. Things I've learned so far: 1. I built way more than I needed to before showing it to anyone. The app wearable integrations, labs, nutrition, recovery, an AI health coach. 2. When you build for everyone, the app ends up complicated and useless for everyone. 3. Feedback from people actually using it taught me more in a week than 2 months of building did. 4. Don't try to build the perfect product. Build the next version that's testable and answers a few questions. Working on that version now.

Just crossed 300 followers :') Started here with nothing. Building Vanta, a health os, out in the open, and most of you are builders doing the same. Small accounts grow together. Let's connect if you're into: - AI agents / vibe coding - iOS / app development - SaaS - Solo founders, building in public - Health, fitness, longevity

3 months of building. the thing that moved the needle most was not the code. it was talking to other builders here. reply with what you're shipping this week, i'll reply to all of them.

Just hit 260 followers. 60 new since yesterday, fastest day so far. Best part is the people. Half my timeline is founders shipping things now and it's the most fun X has ever been. If you're building something, let's connect!

The biggest mistake I made with the first version of my app: I thought in features. People don't stick to features. They stick to seeing themselves get closer to what they want. So I studied what actually makes users stick (Octalysis framework). What matters: 1. Long goals broken into short ones they can hit this week. 2. Visual progress. Every open shows them moving. 3. Triggers on that progress. A notification when they hit a milestone, a nudge when a streak is about to break. 4. Tie progress to their profile, so it becomes theirs and they want to keep building it. 5. Show the one feature they care about, not the 10 features the app has.

app store review took 5 days for a one-line fix. what's the longest you've waited? and did it get rejected?