SIH 2025 - Technical Implementation (Detailed)
This document outlines the technical implementation plan for the SIH 2025 solution.
Each feature includes: What/Why, How, Tech Stack, APIs, Edge Cases, and Ownership.
1) Fully Personalized Experience
- Profile embeddings (skills, interests, quizzes).
- Recommendation engine using hybrid rules + vectors.
- Stack: Next.js, Django+DRF, MySQL, Redis, FAISS/Pinecone.
2) Target Audience Segmentation
- Role-based flows (student/professional/jobseeker).
- Dashboard & content gating via role property.
- Stack: Next.js role-based routes, Django permissions.
3) Mentorship & Guidance (Roadmaps, Resources)
- JSON roadmap schema with drag-drop builder UI.
- Search & discovery via MeiliSearch/Elastic.
- Stack: React DnD, Django+DRF, S3 storage.
4) Skill-based Suggestions & Gap Analysis
- Normalize skill taxonomy & embeddings.
- Gap radar charts in Plotly, roadmap for missing skills.
5) Exam Calendars per User
- Exam DB with official sources, notifications via Celery.
- Calendar export + push/email reminders.
6) Roadmaps with Details
- JSON schema for steps, track progress per user.
- Export to PDF + integrate with calendar.
7) Aptitude & Interest Quizzes
- Configurable quizzes with scoring rules.
- Psychometric scoring → profile enrichment.
8) Course-to-Career Mapping
- Data linking courses → skills → jobs.
- Sankey diagrams in Plotly.
9) Nearby Government Colleges Directory
- Colleges DB with geolocation.
- Leaflet/Mapbox frontend map.
10) Timeline & Notifications
- Celery Beat scheduled reminders (email/push/SMS).
- Dead-letter queue & retries.
11) AI-driven Recommendations
- Hybrid: rules + vectors + collaborative filtering.
- Explainable suggestions with reasons.
12) AI Mentorship (OpenAI GPT)
- Chat wrapper with sanitized prompts.
- Conversation history in DB, token quotas.
- OpenAI moderation API integration.
---
Cross-cutting:
- DB Schema with Users, Quizzes, Courses, Careers, Roadmaps, Resources, Exams, Colleges,
Conversations, Notifications.
- APIs following REST/DRF with JSON responses.
- Background jobs: embeddings, scrapers, notifications.
- Deployment: Vercel (frontend), Railway (backend/workers).
- Security: JWT, HTTPS, PII redaction, moderation.
- Monitoring: Sentry, PostHog analytics.
- Testing: pytest-django, Jest, Cypress.
