AI-Driven SDLC

Business Context

Organizations frequently face the challenge of migrating legacy codebases to modern languages and frameworks. This process is typically manual, error-prone, and requires significant expertise in both source and target technologies. Our AI-Driven SDLC System automates this process using a series of specialized LLM agents that analyze, plan, and execute the migration while keeping humans in the loop for critical decision points.

Problem Statement

Manual code migration is time-consuming, error-prone, and expensive, often requiring specialized knowledge in both legacy and modern systems. Many organizations struggle with modernizing their critical applications due to the complexity and risk involved. This PoC demonstrates how AI-assisted code migration can accelerate the process while maintaining quality and reducing risk.

Impact and Importance

Our AI-Driven SDLC System can reduce migration time by up to 70%, drastically cut costs, and improve the quality of migrated code through consistent patterns and best practices. It enables organizations to modernize faster, reduce technical debt, and improve maintenance and scalability of critical applications.

Developer Setup

To set up and run this PoC locally, follow these steps:

1. Backend Setup

1. Clone the agent repository and install dependencies:

git clone ...

cd ET-AI-AMS-Automation

pip install -r src_v2/requirements.txt

2. Start the Quart server:

python src_v2/app.py

3. Event Stream Format

The event stream sends JSON objects with the following structure:

{
  "event": "update",
  "data": {
    "status": "processing",
    "current_phase": "analyze",
    "files": ["src/main.py", "src/utils.py"],
    "embedding_ready": true,
    "source_info": { ... },
    "analysis_result": { ... },
    // Additional state information
  }
}