cheatcode()

Technical interviews are a pattern recognition game disguised as an intelligence test. We reverse engineered the patterns.

A mobile-first coding challenge app that gets busy engineers interview-ready in 10 minutes a day. Bite-sized problems, real patterns, no fluff.

---

What it is

cheatcode() is the TikTok of Leetcode — short, focused coding challenges served in a daily feed. Instead of grinding full problems for hours, users solve one critical line of code at a time, building pattern recognition through repetition.

Core loop:

1. Open app → see a problem with one blank line
2. Fill in the critical insight
3. Get instant feedback + pattern explanation + brute force vs optimised comparison
4. Next problem

Who it's for: Software engineers preparing for interviews who have 10 minutes a day, not 2 hours.

---

Architecture

┌─────────────────┐     HTTPS      ┌─────────────────┐     SQL      ┌──────────────┐
│   Flutter App   │ ────────────▶  │  FastAPI Backend │ ──────────▶ │  Supabase    │
│   (Android/iOS) │                │  (Vercel)        │             │  (PostgreSQL)│
└─────────────────┘                └─────────────────┘             └──────────────┘

Tech Stack

Layer	Technology
Mobile	Flutter (Dart)
Backend	Python FastAPI
Database	Supabase (PostgreSQL)
Hosting	Vercel (backend)
Auth	Google Sign-In
CI/CD	Vercel auto-deploy on push

---

Project Structure

/
├── cheatcode-ui/          # Flutter mobile app
│   └── lib/
│       ├── main.dart
│       ├── models/
│       │   └── problem.dart
│       ├── providers/
│       │   └── app_provider.dart
│       ├── screens/
│       │   ├── feed_screen.dart       # Main screen — problem attempt
│       │   ├── onboarding_screen.dart
│       │   ├── vault_screen.dart      # Saved EOD problems
│       │   ├── progress_screen.dart   # Pattern ownership
│       │   ├── history_screen.dart    # All solved problems
│       │   └── solved_today_screen.dart
│       ├── services/
│       │   ├── api_service.dart       # Backend API calls
│       │   └── auth_service.dart      # Google Sign-In + session
│       ├── theme/
│       │   └── app_theme.dart
│       └── widgets/
│           ├── dry_run_visualizer.dart
│           └── problem_card.dart
│
└── cheatcode-api/         # FastAPI backend
    ├── main.py             # App entry point + middleware
    ├── config.py           # Environment settings
    ├── Dockerfile
    ├── requirements.txt
    ├── .github/
    │   └── workflows/
    │       └── deploy.yml  # CI/CD pipeline
    ├── db/
    │   └── supabase.py     # Supabase client singleton
    ├── models/
    │   └── problem.py      # Pydantic schemas
    ├── routers/
    │   ├── users.py        # POST /users/, GET /users/{id}
    │   ├── feed.py         # GET /feed/{user_id}
    │   ├── progress.py     # POST /progress/{user_id}
    │   └── answers.py      # POST /answers/{user_id}/validate
    ├── services/
    │   ├── feed_service.py      # Feed algorithm
    │   ├── progress_service.py  # Streak + pattern tracking
    │   └── answer_service.py    # Validation + AI fallback
    └── tests/
        └── test_services.py

---

Database Schema

users               — id, email, name, role, streak, solved_today, interview_date
problems            — id, title, company, pattern, difficulty, code_lines (jsonb),
                      hints (jsonb), wrong_options (jsonb), problem_statement,
                      brute_force, optimised, explanation
patterns            — id, name, description
user_problem_state  — user_id, problem_id, status, hints_used, time_to_solve, solved_at
user_pattern_progress — user_id, pattern_id, times_solved, times_encountered, owned

Problem statuses: unseen → attempted → solved / skipped / vaulted

---

API Reference

Method	Endpoint	Description
POST	/users/	Create or upsert user
GET	/users/{id}	Get user profile
GET	/feed/{user_id}	Get personalised problem feed
POST	/progress/{user_id}	Update problem status
GET	/progress/{user_id}/streak	Get streak + solved today
POST	/answers/{user_id}/validate	Validate answer (with AI fallback)
GET	/answers/{user_id}/solved-today	Get today's solved problems
POST	/answers/{user_id}/unsolved/{problem_id}	Mark problem as unsolved
GET	/health	Health check

Auto-generated docs available at /docs (Swagger) and /redoc.

---

Feed Algorithm

1. Fetch user role + interview date
2. Fetch user's weak patterns (solved < 5 times)
3. Fetch problems already solved/skipped by user
4. Filter by difficulty based on role:
   - student      → [1, 2]
   - professional → [1, 2, 3]
   - competitive  → [2, 3]
5. If interview < 7 days away → bump difficulty to [2, 3]
6. Sort: weak patterns first → difficulty match
7. Fallback: if filter returns empty → serve all active unseen problems

---

Answer Validation

Answers go through two layers:

1. Local normalization — strips whitespace, normalizes operators, lowercases. nums[i-k] and nums[i - k] both match.
2. AI fallback — if normalization fails, asks Claude Haiku whether the answer is logically equivalent. Handles creative but correct solutions.

---

Local Development

Backend

cd cheatcode-api
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env        # fill in Supabase credentials
uvicorn main:app --reload --host 0.0.0.0 --port 8000

API docs: http://localhost:8000/docs

Flutter

cd cheatcode-ui
flutter pub get
flutter run

Important: Update baseUrl in lib/services/api_service.dart:

• Local testing: http://YOUR_MAC_IP:8000 (find with ipconfig getifaddr en0)
• Production: your Vercel URL

For Android physical device, also ensure AndroidManifest.xml has:

android:usesCleartextTraffic="true"

---

Database Setup

1. Create a project at supabase.com
2. Run schema SQL files in order in SQL Editor:
   • query_seed_30_fixed.sql — patterns + 30 problems
   • query_problem_statements.sql — problem descriptions
3. Copy Project URL and service_role key to .env

---

Deployment

Backend (Vercel)

1. Create a new project at vercel.com from this repository.
2. Set Root Directory to cheatcode-api.
3. Set environment variables in Vercel:

   SUPABASE_URL=...
   SUPABASE_SERVICE_KEY=...
   APP_ENV=production
   API_SECRET_KEY=...   # required by middleware in production
   

4. Deploy. vercel.json routes all paths to the FastAPI ASGI app.

Flutter

# Android release build
flutter build apk --release

# Upload to Google Play Console

---

Content

30 problems across 5 patterns, mapped to Neetcode 150:

Pattern	Problems	Companies
Sliding Window	6	Amazon, Google, Meta, Microsoft
Two Pointer	6	Amazon, Google, Meta, Facebook
Binary Search	6	Google, Amazon, Meta
Hash Map	6	Google, Meta, Amazon
Dynamic Programming	6	Amazon, Google

Each problem has:

• Problem statement
• Code with one critical blank
• 3 progressive hints
• 4 multiple choice options (1 correct + 3 plausible wrong answers from DB)
• Pattern explanation
• Brute force vs optimised comparison with complexity
• Interactive dry run visualizer

---

Roadmap

• Push notifications — daily challenge reminder
• Friend leaderboard — weekly solved count
• Company-specific problem sets
• Interview timeline mode — auto-schedule by date
• Community problem submissions
• Paywall — Pro tier (unlimited hints, vault, company sets)
• Web version

---

Environment Variables

Variable	Description
SUPABASE_URL	Supabase project URL
SUPABASE_SERVICE_KEY	Supabase service role key (never expose client-side)
APP_ENV	development or production

---

Running Tests

cd cheatcode-api
pytest tests/ -v

Tests cover streak logic, feed algorithm edge cases, and answer normalization.

---

Why "cheatcode"

Interviews test pattern recognition, not raw intelligence. Every technical interview draws from a pool of ~20 fundamental patterns. Once you internalize those patterns, you recognize them regardless of how the problem is disguised.

That's not cheating. That's preparation. But it feels illegal.

---