Shrish Kerur
AI Developer | Generative AI & Agentic Systems | Python
Introduction
I am a developer building AI applications and LLM-based systems. I began with full-stack development during my BCA and have since moved into Machine Learning and the LangChain ecosystem during my MCA. I use statistics and mathematics to build and rigorously evaluate my projects. I am currently focused on agentic workflows and model fine-tuning.
Education
Master of Computer Applications (MCA)
Bangalore Institute Of Technology
2024 — Present
Bachelor of Computer Applications (BCA)
Basaveshwara Science College
2021 — 2024
Skills
AI & ML — LangChain, LangGraph, RAG, Agentic Workflows, DeepEval, PyTorch, Keras, Scikit-learn
Data & Math — Python, NumPy, Pandas, Statistics, Linear Algebra, Calculus
Databases — Qdrant, ChromaDB, PostgreSQL, MongoDB
Tools — Git, Docker, Linux, Groq, HuggingFace
Recent Projects
Synthetix — Autonomous Git Agent
[SOURCE]Python · LangGraph · Groq · PyGithub
An agentic AI built with LangGraph that accepts a GitHub repository URL and autonomously refactors all programming files. Uses a state machine with two Groq-powered LLMs — one for refactoring and one for review — with a self-correction loop that re-refactors any file scoring below 0.5. Once all files are processed, it raises a pull request to the main branch.
LegalRAG — Document Audit Tool
[SOURCE]Python · LangChain · Chroma · FlashRank · DeepEval · Docker
A RAG application for analysing legal documents using Qdrant for vector storage and FlashRank for re-ranking retrieved chunks. Evaluated using DeepEval across four RAG metrics — all scoring above 0.8.
EduRAG — Assessment Generator
[SOURCE]Python · LangChain · ChromaDB · Google Vision · Tesseract · Gemini
A tool that processes syllabus PDFs to extract modules, course outcomes, and Bloom's Taxonomy levels. Generates question papers based on user-specified format and difficulty. Students can upload a photo of handwritten answers and receive LLM-powered feedback via Google Vision and Gemini.