⚡ Python AI Backend Architecture (Capstone)

A production-grade implementation of asynchronous Python concepts, designed to mock and manage high-concurrency LLM streaming infrastructure.

🧠 Architecture Overview

This repository demonstrates how to bridge the gap between standard Python scripting and enterprise AI backend engineering. Moving from a MERN stack background, this project enforces:

• Strict Concurrency: Using Python 3.11+ asyncio.TaskGroup to manage parallel I/O bound LLM requests without blocking the Event Loop.
• Resilience: Custom @async_retry decorators implementing Exponential Backoff and Jitter to prevent Thundering Herd problems during API outages.
• Telemetry: Custom @async_timer decorators leveraging functools.wraps to preserve function metadata for AI observability tools (like LangSmith).
• Streaming: Utilizing Asynchronous Generators (yield) to simulate Server-Sent Events (SSE) for token-by-token LLM output.
• Type Safety: Pydantic models for strict data validation on streamed chunks.

🚀 Execution

Developed using uv for lightning-fast dependency management.

🛠️ Prerequisites

Make sure you have Python 3.11+ and uv installed.

If you don't have uv installed, get it via:

# macOS/Linux
curl -LsSf [https://astral.sh/uv/install.sh](https://astral.sh/uv/install.sh) | sh

# Windows
powershell -c "irm [https://astral.sh/uv/install.ps1](https://astral.sh/uv/install.ps1) | iex"

🚀 How to Run

1. Clone the repository:

git clone https://github.com/Piyushtiwari919/Python-backend-lab.git

cd Python-backend-lab

2. Install dependencies:

uv sync

3. Run the application:

uv run python -m src.main

🏗️ Next Steps & Integration

This backend architecture acts as the foundational layer for AI agent workflows. The concepts validated here—specifically non-blocking generator streams and strict task lifecycle management—are designed to be integrated seamlessly with asynchronous WebSocket clients and React/Next.js frontends.