›_
Back to Projects
Cognity-AI
in-progressMar 2026 - Present

Cognity-AI

Personal

Working on a Python library that provides a unified interface for multiple RAG (Retrieval-Augmented Generation) methodologies and LLM providers. The goal is to drastically reduce the boilerplate needed to configure and switch between different RAG pipelines, embedding models, and LLM backends — enabling rapid experimentation and production deployment with minimal code.

Tech Stack

PythonRAGLangChainLlamaIndexLLMVector DatabasesPyPI

Highlights

  • Unified interface for multiple RAG strategies
  • Pluggable LLM provider support
  • Minimal-code pipeline configuration

Challenges

  • Designing a consistent API that works across fundamentally different RAG architectures (naive, modular, agentic, graph-based)
  • Abstracting provider-specific quirks (rate limits, token formats, streaming differences) without leaking complexity
Know More Source Code