Skip to content
View SandeshGitHub2077's full-sized avatar

Block or report SandeshGitHub2077

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
SandeshGitHub2077/README.md

Hi there 👋 I'm Sandesh Mahajan

AI/ML Engineer | Generative AI | Agentic Workflows | Graph RAG | Document Intelligence

I build production-grade AI systems for enterprise workflows, document intelligence, retrieval-augmented generation, and automation.


About Me

I am an AI/ML Engineer specializing in Generative AI, Agentic Workflows, and Retrieval-Augmented Generation (RAG) systems.

My work focuses on building scalable AI applications that combine LLMs, knowledge graphs, vector search, OCR, structured validation, and cloud-native deployment. I enjoy turning complex unstructured documents into searchable, auditable, and production-ready AI systems.

Currently, I work on enterprise AI systems involving:

  • Graph RAG across large-scale PDFs and image-based documents
  • OCR-driven document intelligence pipelines
  • LangChain-based AI orchestration workflows
  • Neo4j knowledge graphs and PostgreSQL metadata indexing
  • Schema-constrained LLM outputs using Pydantic and JSONSchema
  • Secure AI workflows with audit logging, validation, and CI/CD deployment

Stuff I Excel At

Generative AI & Agentic Systems

RAG Badge Graph RAG Badge Agentic Workflows Badge LangChain Badge Prompt Engineering Badge

Vector Search & Knowledge Systems

FAISS Badge Pinecone Badge ChromaDB Badge Neo4j Badge Hybrid Search Badge

Languages

Python  Java  R  C++ 

Machine Learning & NLP

PyTorch  TensorFlow  Scikit-learn Badge Transformers Badge BERT Badge

Cloud & Infrastructure

AWS  Azure  Docker  Kubernetes  GitHub Actions 

Databases

PostgreSQL  MySQL  SQL Server Badge Vector DB Badge Neo4j Badge

Developer Tools

Git  GitHub  VS Code  Jupyter 

Cursor Badge Claude Code Badge OpenAI Codex Badge OpenClaw Badge GitHub Copilot Badge

Featured Work

Enterprise Graph RAG System

Built an enterprise Graph RAG system for large-scale PDFs and image-based documents using OCR, hierarchy-aware chunking, embeddings, Neo4j knowledge graphs, LangChain orchestration, and PostgreSQL metadata indexing.

Core capabilities:

  • Hybrid retrieval using graph traversal, vector embeddings, and BM25 ranking
  • Parent-child and sibling document relationships for contextual reasoning
  • Cross-document retrieval over deeply nested enterprise content
  • Schema-constrained LLM outputs using Pydantic and JSONSchema
  • Audit-ready validation, logging, and deterministic output enforcement

OCR-Driven Document Intelligence Platform

Developed an OCR-driven AI workflow platform using Azure AI Foundry OCR, FastAPI, React, SQL Server, and XML-based integrations.

Core capabilities:

  • Automated document processing across PDFs, images, and email attachments
  • OCR extraction and validation pipelines
  • Role-based access control
  • API submission workflows
  • Audit logging for secure enterprise AI operations
  • Markdown and machine-readable output generation for downstream AI workflows

Agentic Workflow Automation

Designed LLM-driven automation workflows for lead qualification, campaign routing, decision logic, and business-rule execution.

Core capabilities:

  • Multi-step autonomous workflow orchestration
  • Prospect classification and redistribution
  • Conditional routing across enterprise workflow systems
  • Reduced manual intervention through AI-driven automation

Currently Exploring

  • Production-grade RAG evaluation
  • Graph-based retrieval optimization
  • Agentic AI workflow design
  • OCR + LLM document transformation
  • Secure AI systems for regulated environments
  • Cloud-native AI deployment patterns

Connect With Me

LinkedIn Badge   Gmail Badge   GitHub Badge

Pinned Loading

  1. LegalClaimGPT LegalClaimGPT Public

    AI-powered legal assistant that predicts personal injury settlement amounts from case summaries using LLMs, RAG, and explainable ML. Built with FastAPI, Streamlit, and XGBoost.

    Python

  2. transit_optimizer_ai transit_optimizer_ai Public

    Transit Optimization Assistant using AI + GTFS A smart geospatial tool that helps transit agencies and NGOs improve public transport by analyzing GTFS bus stop data. It detects redundant stops, fin…

    Python

  3. autogen_multiagent_groupChat autogen_multiagent_groupChat Public

    A collaborative AI workflow using Microsoft AutoGen to scrape, categorize, and summarize recent LLM-related arXiv papers with agents like Planner, Engineer, Scientist, and Executor. All automatical…

    Jupyter Notebook

  4. Doctor-Appointment-Application Doctor-Appointment-Application Public

    A scalable microservices-based appointment booking system developed with Java Spring Boot, Docker, and Kubernetes. This system streamlines patient-doctor scheduling, admin workflows, and feedback c…

  5. agentic-hybrid-rag agentic-hybrid-rag Public

    Agentic hybrid RAG system using FAISS, Neo4j, LangGraph, and Ollama for intelligent knowledge retrieval and reasoning.

    Python

  6. multi-hop-graphrag multi-hop-graphrag Public

    An end-to-end GraphRAG system that parses structured documents (RFCs), extracts intra- and inter-document references, builds a Neo4j knowledge graph, and performs hybrid retrieval using FAISS and g…

    Python