CareerDojo

• 

  Interview

• 

  Practice

• 

  Dashboard

• 

  Roadmap

Inspiration

Millions of students in India prepare for competitive exams and tech interviews every year — UPSC, SSC, FAANG, NDA, Banking, MBA. But preparation is scattered across multiple apps, YouTube channels, and coaching notes. There was no single intelligent platform that adapts to YOUR specific goal and generates real practice content on demand. CareerDojo was built to change that.

What it does

CareerDojo is an AI-powered career preparation platform with three core features:

1. Practice Arena — Topic-wise MCQ practice for your specific goal. FAANG candidates get DSA and System Design questions. UPSC aspirants get Polity, History, Geography in authentic statement-based format. SSC, NDA, Banking and MBA students get questions matching their exact exam pattern. All questions are generated fresh by AI based on real exam patterns.

2. Mock Interviews — AI conducts a full interview tailored to your goal. FAANG gets Technical and Behavioral rounds. UPSC gets Personality Test simulation. NDA gets SSB-style questions. After each answer the AI scores you, highlights strengths, suggests improvements and shows a model answer.

3. My Roadmap — A personalized 14-day study plan generated based on your goal, skill level, daily available hours and target date. Updates based on your practice performance. Track daily progress with checkboxes and a progress bar.

Bonus features: Skill Gap Analysis with color coded charts, daily streaks, badges, daily motivational tips and a profile with detailed statistics.

How I built it

I built CareerDojo entirely using MeDo — Baidu's AI-powered no-code app builder. I described every feature, database schema, UI design and AI behavior in natural language through multi-turn conversations with MeDo. MeDo generated the full-stack application including frontend, backend Edge Functions on Supabase, database tables and Gemini AI integration — all without writing a single line of code manually.

The most impressive part MeDo built was the goal-based AI routing system — a single universal prompt that automatically adapts to any goal the user selects, whether FAANG, UPSC, SSC, NDA, Banking, Railway or MBA, and generates authentic exam-specific content every time.

Challenges I ran into

The biggest challenge was getting the Gemini API responses to parse correctly inside MeDo's Edge Functions. Gemini sometimes returns backtick characters or markdown fences inside JSON which breaks standard parsers. I solved this by implementing a 4-strategy bulletproof parser and increasing maxOutputTokens to prevent response truncation.

Another challenge was making the roadmap persist correctly — the app was deleting the old roadmap before saving the new one, leaving users with nothing if saving failed. Fixed by reversing the save-then-delete order.

Accomplishments

• Built a complete full-stack AI application with zero manual coding using only MeDo
• Covers 10+ career goals with authentic exam-specific content for each
• Universal AI prompt system that works for any goal without hardcoding
• Fully functional on both desktop and mobile

What I learned

MeDo is genuinely powerful for building real applications fast. The key insight is that the quality of what MeDo builds depends entirely on how clearly and specifically you describe what you want. Detailed, structured natural language prompts produce professional full-stack applications.

What's next

• Voice-based mock interviews
• Peer comparison and leaderboards
• Integration with job portals
• Native mobile app
• Premium subscription with advanced analytics

Built With

• css
• google-gemini-ai-api
• html
• javascript
• medo
• supabase