CaseForge

AI-Powered Case Study Generator for Professional Development

Transform how students learn business strategy through dynamic, AI-generated case studies with intelligent evaluation and personalized feedback.

Overview

CaseForge is an intelligent case study generation platform built for educational institutions and corporate training. It uses LangGraph state machines and LLMs to create unique, realistic business scenarios on-demand—no templates, no repeats.

Current Status: Phase 2 - LangGraph + Tools ✅

Features

For Students

Dynamic Case Generation - Unique cases generated every time
Multi-Level Difficulty - Beginner, Intermediate, Advanced
Intelligent Evaluation - AI-powered solution scoring
Personalized Feedback - Actionable insights from AI mentor
Case History - Track your learning journey
Industry Variety - FinTech, Healthcare, E-commerce, SaaS, and more

For Institutions

Scalable to 1000+ students - Built for colleges
No Content Management - Cases generate automatically
Progress Tracking - Monitor student performance
API-First - Integrate with existing LMS platforms
Real Data Tools - Market research, financial analysis, competitive intel
Enterprise-Ready - Production-grade infrastructure

Architecture

┌─────────────────────────────────────────────────────────┐
│                    FastAPI Server                       │
├─────────────────────────────────────────────────────────┤
│
├─ REST API Routes
│  ├─ POST /api/v1/cases/generate
│  ├─ POST /api/v1/solutions/evaluate
│  ├─ GET /api/v1/cases/{case_id}
│  └─ GET /api/v1/users/{user_id}/cases
│
├─ LangGraph State Machine (Workflow)
│  ├─ Node: generate_case (Groq LLM)
│  ├─ Node: validate_case (Quality checks)
│  ├─ Node: refine_case (Auto-improve if needed)
│  └─ Node: save_case (Database persistence)
│
├─ Services & Tools
│  ├─ GroqService (LLM API wrapper)
│  ├─ WorkflowService (LangGraph executor)
│  ├─ CaseService (Business logic)
│  └─ Tools (Market research, Financial analysis, Competitive intel)
│
└─ Database (SQLAlchemy + SQLite)
   ├─ case_studies (Generated cases)
   ├─ user_solutions (Student submissions)
   └─ users (User profiles)

Tech Stack

Component	Technology
Framework	FastAPI 0.104
Agentic AI	LangGraph 0.0.15
LLM	Groq (llama-3.3-70b-versatile)
Database	SQLAlchemy + SQLite
Language	Python 3.12
Async	asyncio, uvicorn

📋 Project Structure

caseforge/
├── main.py                          # FastAPI entry point
├── config.py                        # Settings from .env
├── graph.py                         # LangGraph state machine
├── requirements.txt                 # Dependencies
│
├── app/
│   ├── api/
│   │   └── routes.py               # REST endpoints
│   │
│   ├── services/
│   │   ├── groq.py                 # Groq API wrapper
│   │   ├── case.py                 # Case generation logic
│   │   └── workflow.py             # LangGraph executor
│   │
│   ├── workflows/
│   │   ├── state.py                # State definition
│   │   └── nodes.py                # Workflow nodes
│   │
│   ├── tools.py                    # Market research, financial analysis, etc.
│   ├── prompts.py                  # LLM prompts
│   ├── models.py                   # SQLAlchemy models
│   ├── db.py                       # Database setup
│   └── logger.py                   # Logging
│
├── scripts/
│   └── init_db.py                  # One-time DB initialization
│
└── caseforge.db                    # SQLite database

🔧 Installation & Setup

Prerequisites

Python 3.12+
Groq API key (free at https://console.groq.com)

1. Clone & Setup

git clone <your-repo>
cd caseforge
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

2. Install Dependencies

pip install -r requirements.txt

3. Configure Environment

cp .env.example .env
# Edit .env with your Groq API key
GROQ_API_KEY=gsk_your_key_here

4. Initialize Database

python scripts/init_db.py
# Or let the app initialize on startup

5. Run the Server

uvicorn main:app --reload

Server runs at http://localhost:8000 API docs at http://localhost:8000/docs (Swagger UI)

API Usage

Generate a Case Study

Request:

curl -X POST http://localhost:8000/api/v1/cases/generate \
  -H "Content-Type: application/json" \
  -d '{
    "user_id": 1,
    "industry": "FinTech",
    "complexity": "beginner",
    "focus_area": "Product Strategy",
    "time_limit": 60
  }'

Response:

{
  "success": true,
  "case_id": 2,
  "case_uuid": "420c20a7-2eff-4218-b0d8-afae5f0578d2",
  "title": "FinClarity Case Study",
  "industry": "FinTech",
  "complexity": "beginner",
  "case_data": {
    "company_name": "FinClarity",
    "scenario_overview": "...",
    "key_metrics": {...},
    "discussion_questions": [...],
    "hidden_assumptions": [...],
    "solution_framework": "..."
  },
  "generation_time_ms": 2969,
  "refinements_used": 0
}

Evaluate a Solution

Request:

curl -X POST http://localhost:8000/api/v1/solutions/evaluate \
  -H "Content-Type: application/json" \
  -d '{
    "user_id": 1,
    "case_id": 2,
    "solution": "Your proposed solution text here..."
  }'

Response:

{
  "success": true,
  "solution_id": 1,
  "scores": {
    "overall": 8,
    "problem_understanding": 9,
    "analytical_rigor": 7,
    "business_acumen": 8,
    "communication": 8,
    "feasibility": 7
  },
  "feedback": {
    "strengths": ["Clear problem identification", "Data-driven approach"],
    "weaknesses": ["Missed competitive positioning"],
    "suggestions": ["Consider market trends", "Model financial impact"]
  }
}

Get User's Case History

Request:

curl http://localhost:8000/api/v1/users/1/cases

Response:

{
  "success": true,
  "user_id": 1,
  "total": 2,
  "cases": [
    {
      "id": 2,
      "uuid": "420c20a7-...",
      "title": "FinClarity Case Study",
      "industry": "FinTech",
      "complexity": "beginner",
      "created_at": "2026-05-11T02:04:52"
    },
    ...
  ]
}

LangGraph Workflow

The heart of CaseForge is a state machine that ensures high-quality case generation:

START
  ↓
[GENERATE] - LLM creates raw case
  ↓
[VALIDATE] - Check quality & completeness
  ↓
  ├─ ✅ Valid? → [SAVE] → END
  │
  ├─ ❌ Invalid & retries left? → [REFINE] → back to VALIDATE
  │
  └─ ❌ Max retries? → [ERROR] → END

Features:

✅ Automatic validation of case structure
✅ Self-healing - refines bad cases automatically
✅ Prevents invalid cases from being saved
✅ Full audit trail in logs
✅ Sub-3 second generation time

📊 Database Schema

case_studies

CREATE TABLE case_studies (
  id INTEGER PRIMARY KEY,
  uuid VARCHAR(36) UNIQUE,
  user_id INTEGER,
  title VARCHAR(200),
  industry VARCHAR(100),
  complexity ENUM(beginner, intermediate, advanced),
  focus_area VARCHAR(200),
  case_data JSON,
  generation_time_ms INTEGER,
  tokens_used INTEGER,
  model_used VARCHAR(100),
  refinement_count INTEGER,
  created_at DATETIME
);

user_solutions

CREATE TABLE user_solutions (
  id INTEGER PRIMARY KEY,
  uuid VARCHAR(36) UNIQUE,
  user_id INTEGER,
  case_id INTEGER,
  solution_text VARCHAR(5000),
  overall_score FLOAT,
  reasoning_score FLOAT,
  communication_score FLOAT,
  business_acumen_score FLOAT,
  feedback_data JSON,
  created_at DATETIME
);

Tools Available

Market Research Tool

CaseStudyTools.market_research(industry, company_type)
# Returns: market size, growth rate, trends, competitors

Financial Analysis Tool

CaseStudyTools.financial_analysis(industry, user_count, arpu)
# Returns: CAC, LTV, payback period, projections

Competitive Intelligence Tool

CaseStudyTools.competitive_intelligence(industry, segment)
# Returns: competitor analysis, white space, defensibility

Metrics & Performance

Metric	Value
Case Generation Time	~2-4 seconds
Validation Success Rate	~95% (first try)
Average Refinements	0.1 per case
API Response Time (p95)	<5 seconds
Database Queries/Case	3-5
Scalability Target	1000+ concurrent students

Security

✅ Environment variables for secrets
✅ Input validation (Pydantic)
✅ SQL injection prevention (SQLAlchemy)
✅ Rate limiting ready (middleware)
✅ Error handling without exposing internals
✅ Logging for audit trails

Development Roadmap

Phase 1 ✅ (May 11)

✅ FastAPI server
✅ Groq LLM integration
✅ Basic case generation
✅ Solution evaluation
✅ SQLite database

Phase 2 ✅ (May 12)

✅ LangGraph state machine
✅ Validation + refinement workflow
✅ Tools (market research, financial analysis)
✅ Competitive intelligence

Phase 3 (May 13-15)

Authentication (JWT)
Difficulty progression (adaptive)
WebSocket streaming
Progress dashboard

Phase 4 (May 16-24)

Deployment to production
Load testing (1000+ concurrent)
Monitoring & alerting
Documentation & tutorials

Phase 5 (May 25)

Final demo
Performance optimization
Feedback incorporation

Contributing

Create a feature branch
Make your changes
Test locally
Commit with clear messages
Push and create PR

License

Proprietary - For educational use only

Author

Built for college students by Aniketh Cheerath

Contact: cheerathaniketh@gmail.com

Support

Common Issues

Q: Groq API key not working?

Check .env file has GROQ_API_KEY=...
Verify API key is active at https://console.groq.com
Restart server after changing .env

Q: Cases are too generic?

Tools are integrated - they provide real market data
Prompts are continuously improved
Each case is unique - never repeats

Q: How to scale to 1000+ students?

Database is async-ready
Use PostgreSQL instead of SQLite for production
Add Redis caching for prompts
Use load balancer + multiple server instances
Consider Ray for distributed LLM calls

Resources

Made with ❤️ for education

Name		Name	Last commit message	Last commit date
Latest commit History 17 Commits
app		app
scripts		scripts
view		view
.gitignore		.gitignore
README.md		README.md
config.py		config.py
graph.py		graph.py
main.py		main.py
requirements.txt		requirements.txt
submit_solution.py		submit_solution.py

Folders and files

Latest commit

History

Repository files navigation

CaseForge

Overview

Features

For Students

For Institutions

Architecture

Tech Stack

📋 Project Structure

🔧 Installation & Setup

Prerequisites

1. Clone & Setup

2. Install Dependencies

3. Configure Environment

4. Initialize Database

5. Run the Server

API Usage

Generate a Case Study

Evaluate a Solution

Get User's Case History

LangGraph Workflow

📊 Database Schema

case_studies

user_solutions

Tools Available

Market Research Tool

Financial Analysis Tool

Competitive Intelligence Tool

Metrics & Performance

Security

Development Roadmap

Phase 1 ✅ (May 11)

Phase 2 ✅ (May 12)

Phase 3 (May 13-15)

Phase 4 (May 16-24)

Phase 5 (May 25)

Contributing

License

Author

Support

Common Issues

Resources

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Packages