How This Ex-Trader Built a $15K/Month App Portfolio Using Cursor AI

TL;DR: Key Takeaways

David Attias spent 7 years designing algorithmic trading systems for banks. In 2025, he watched three teenagers build a $200K/month app (QUITTR) that was objectively terrible — and thought, "If they can do it, I can do better." One year later, he's earning $15,000/month from a portfolio of mobile apps, built almost entirely with Cursor AI and spending just $200/month on AI tools.

Core takeaway: In 2026, the bottleneck isn't building — it's distribution. David's genius wasn't in code (he used Cursor to vibe code everything). It was in his influencer equity model: give big creators a stake in your app instead of paying per post, and suddenly they have skin in the game.

Who Is David Attias?

David graduated from a computer science school in Paris 13 years ago. He started his career at Criteo as a technical account manager, working across London, Barcelona, and New York. But his real education came from the next seven years: designing trading algorithms for banks and professional traders.

Algorithmic trading teaches you one thing: precision matters, but speed matters more. A strategy that's 80% right today beats one that's 100% right next month. David carried this mindset into indie hacking.

"Even a year ago, AI was 70-80% perfect for mobile app development. I still had to test each screen, button, navigation, and flow. But whenever I spotted an error, I just asked Cursor to fix it." — David Attias, Indie Hackers interview

The "Aha" Moment: Why Build When You Can Copy?

In 2025, David watched a podcast featuring three teenagers who built QUITTR, a smoking-cessation app that hit $200,000 per month in just three months. He downloaded the app, paid for it, and was stunned by its poor quality.

This was the insight that changed everything: a bad product was succeeding because it solved a real problem and had killer distribution. The market had already been validated — all David needed to do was execute better.

He built STOPPR, an app to help people break processed sugar cravings, modeled on QUITTR's onboarding flow but with better UX. Within two weeks, two viral influencer videos pushed it to $5,000 in revenue. Three months later, it hit $14,000/month.

Step-by-Step: How David Built a $15K/Month App Portfolio

Step 1: Validate Demand by Copying What Already Works

Don't invent. Improve. David's first move wasn't to brainstorm — it was to find a proven winner and ask: "Can I do this better?"

Key principle: You don't need to validate market demand from scratch. If a bad product is making money, the demand is real. Your job is execution, not discovery.

Step 2: Set Up the AI Development Pipeline

David's original stack (2025):

• Cursor — AI-powered IDE for vibe coding
• Figma — Design tool, imported into Cursor via Figma MCP
• Claude 3.5 — AI model for code generation
• Firebase — Backend (auth, database, hosting)

The Figma MCP workflow was the game-changer:

1. Design screens in Figma (David hired a designer friend for this)
2. Import designs directly into Cursor via the Figma MCP integration
3. Ask Cursor to "vibe code" both frontend and backend for each screen
4. Test each screen, button, navigation, and flow manually
5. When errors appear, ask Cursor to fix them

"The barrier to design and code was incredibly low then. It's even lower now." — David Attias

David's current stack (2026):

• GPT 5.5 CLI within Cursor (replaced Claude)
• Firebase (unchanged)
• Monthly AI cost: $200 (down from $1,000)

His verdict on the switch: "GPT 5.5 is goated for mobile app dev. Better than Claude 4.7."

Step 3: Ship the MVP in 2 Weeks

David's speed comes from a simple rule: the first version doesn't need to be perfect — it needs to exist.

His 2-week sprint looked like this:

Critical insight: David spent ZERO time on "perfect code." He let Cursor generate everything, tested the user experience, and only fixed what was broken. This is the essence of vibe coding: the user doesn't care about your code quality — they care about whether the app works.

Step 4: The Distribution Playbook (This Is Where the Money Is)

David's distribution strategy evolved through two phases:

Phase 1 (First 3 months) — Pay Per Post

• Pay influencers 20% upfront for 8 videos/month (4 Instagram + 4 TikTok)
• Requirement: hit cumulative minimum views based on their last 20 video averages
• Result: $5K revenue in 2 weeks, $14K/month at 3 months
• Problem: Influencers lacked incentive to drive conversions — they got paid regardless

Phase 2 (Current) — Influencer Equity Model

• Find 1-2 big influencers in the niche (not UGC creators, not micro-influencers)
• Offer them equity in the app instead of cash payment
• Align incentives: they only win when you win
• Amplify virality: when a video starts gaining traction, immediately boost it with TikTok and Meta paid ads via viral.app

"Most influencers actually lack influence. They don't know how to go viral on repeat — that's a learnable skill. I only need one or two per app." — David Attias

How he finds them: Directly through the "For You" feed on TikTok and Instagram. No creator marketplace platforms ("They're overpriced").

Step 5: Optimize and Scale Across Multiple Apps

Once STOPPR was generating consistent revenue, David didn't stop. He applied the same playbook to additional apps, treating his portfolio like a system rather than individual bets.

The portfolio approach:

• Reuse the same tech stack (Cursor + Firebase)
• Reuse the same distribution playbook (influencer equity)
• Reuse the same monetization patterns (subscription-based)
• Learn from each app's failures and apply to the next

David now earns roughly $15,000/month: 70% from his app portfolio, 30% from his day job as Developer Relations Advocate at Adapty.

Revenue Breakdown

Cost structure:

Profit margin on apps alone: ~96% (not counting influencer equity dilution)

Tool Cost Comparison: Then vs Now

This is one of the most striking data points from David's journey — the cost of AI tools dropped 80% while quality improved:

The lesson: AI development costs are in freefall. What cost $1,000/month in 2025 costs $200/month in 2026 — and the output is better.

3 Key Insights From David's Journey

Insight 1: Distribution Is the Real Moat, Not Code

David was a professional algorithm designer with 7 years of experience. He could have written perfect, production-grade code. Instead, he let Cursor handle the coding and focused all his energy on finding and converting users.

In 2026, anyone can build an app with AI. Very few people know how to get it in front of users. The value isn't in what you build — it's in who sees it and why they pay.

Insight 2: Equity Over Cash Aligns Everyone

The switch from paying per post to offering equity was David's highest-leverage decision. When influencers have skin in the game:

• They push harder for conversions (not just views)
• They stay engaged long-term (they own a piece)
• They filter themselves (only those who believe join)

This is the same logic behind startup employee stock options, applied to influencer marketing.

Insight 3: The "Bad Product" Test Is the Best Validation

David didn't do surveys. He didn't run ads to test demand. He found a product that was objectively bad but financially successful — and that was all the validation he needed.

The "Bad Product" Test:

1. Find a successful app in your target niche
2. Use it yourself. Pay for it.
3. If it sucks but people still pay → validated market
4. Build a better version

Lessons Learned & Failures

What didn't work:

• UGC and ambassador programs: "They don't know how to go viral. Micromanaging them takes too much time."
• Paying per post: Created zero incentive alignment; influencers optimized for views, not conversions
• Trying to manage 8+ influencers per app: Too much overhead; 1-2 big ones is the sweet spot

What David would do differently:

• Start with the equity model from day one (not waste 3 months on pay-per-post)
• Go straight to GPT 5.5 CLI (skip Claude for mobile dev)
• Build the influencer relationship before launching the app

Can You Do This Too?

David's path is replicable, but it requires shifting your mindset from "I need to code this perfectly" to "I need to ship this and find users."

If you're starting from zero:

1. Find a bad-but-profitable app in a niche you understand
2. Set up Cursor + Figma MCP + Firebase (cost: ~$200/month)
3. Ship an MVP in 2 weeks — don't aim for perfection
4. Find 1-2 influencers in the niche via TikTok/Instagram For You feed
5. Offer equity, not cash — align incentives from day one
6. When a video goes viral, pour paid fuel on it immediately

If you already have an app but growth is stuck:

• The problem is probably distribution, not product
• Switch from pay-per-post to equity-based influencer deals
• Use viral.app or similar to track which content is taking off
• Boost winning content with paid ads within hours (not days)

🔒 Pro Version Includes

The full Pro version of this case study includes:

• Complete 7-step execution blueprint with Cursor configuration files
• Revenue verification data from App Store Connect and Firebase analytics
• Detailed influencer outreach templates (DMs, emails, call scripts)
• Equity agreement template for influencer partnerships
• Figma → Cursor MCP setup guide with exact CLI commands
• 3 common failure modes and how to avoid them (influencer ghosting, app rejection, paid ad waste)

📚 Learn Cursor

• Cursor 2.0 & Composer 2 Guide: Setup, Pricing, Benchmarks (2026)

📖 Related Cases

• Full-Stack Dev Built Real Estate Lead Scoring System with n8n + Claude: $3,500/Month
• How He Built an AI Automation System with Claude + n8n: $4K to $12K/Month in 6 Months
• Claude Code 48-Hour Startup: One Person + $29/Month, $9K MRR in 3 Months

Source: Indie Hackers — "From zero to $10k/mo app portfolio in a year" by James Fleischmann, May 22, 2026. Interview with David Attias, founder of STOPPR.