Building intelligent
systems across
the cosmos of data.

Senior GenAI Engineer with 3+ years of experience specializing in AI Agents, RAG, enterprise automation, and LLM-powered applications — shipping production-grade AI for document intelligence, automation and analytics.

Mumbai, Maharashtra
Shraddha Abhang
AI Agents Multi-Agent Systems LangGraph LangChain LlamaIndex RAG Google ADK Vertex AI Vertex AI RAG Engine Gemini OpenAI GPT-4o Ollama LangFlow MCP Tool Calling Guardrails LLM Evaluation Prompt Caching OpenTelemetry Prompt Engineering Text-to-SQL Vanna AI FAISS Qdrant Neo4j ChromaDB SpaCy Surya OCR PyTesseract PyMuPDF Computer Vision YOLO OpenCV FastAPI SSE Python Docker Kubernetes Redis PostgreSQL MongoDB Cloud SQL AWS Bedrock GCP Azure Streamlit Gradio React Next.js TensorFlow PyTorch XGBoost Unleash AI Agents Multi-Agent Systems LangGraph LangChain LlamaIndex RAG Google ADK Vertex AI Vertex AI RAG Engine Gemini OpenAI GPT-4o Ollama LangFlow MCP Tool Calling Guardrails LLM Evaluation Prompt Caching OpenTelemetry Prompt Engineering Text-to-SQL Vanna AI FAISS Qdrant Neo4j ChromaDB SpaCy Surya OCR PyTesseract PyMuPDF Computer Vision YOLO OpenCV FastAPI SSE Python Docker Kubernetes Redis PostgreSQL MongoDB Cloud SQL AWS Bedrock GCP Azure Streamlit Gradio React Next.js TensorFlow PyTorch XGBoost Unleash
97%
Document processing time reduced
2.8M+
Emails automated every year
3+
Years in Applied GenAI & ML
1
Published research paper
Saturn
// about

I turn research notebooks into AI products you can actually ship.

Currently Senior GenAI Engineer at Indian Tec Solutions, specializing in AI Agents, multi-agent systems, RAG pipelines, and enterprise automation. Building document intelligence platforms, automated banking & KYC workflows, and LLM evaluation frameworks. Published researcher with a track record of transforming business problems into production-grade AI systems.

Generative AI

RAG pipelines, LLM apps, LangChain, LlamaIndex, Ollama, OpenAI.

Computer Vision

OCR, YOLO, OpenCV — document intelligence at production scale.

ML Engineering

TensorFlow, XGBoost, classification & predictive analytics.

Cloud & Deploy

AWS EC2/S3, Azure, Docker, FastAPI services in UAT & Prod.

// experience

A timeline of shipped AI.

Dec 2025 — Present

Senior Gen AI Engineer

Indian Tec Solutions
  • Architected and developed a multi-agent customer support platform using Gemini, Google ADK, and RAG, automating issue resolution, ticket creation, and website assistance.
  • Implemented enterprise-grade guardrails, prompt caching, and AI safety controls, reducing inference costs by 75% while improving response quality and reliability.
  • Designed RAG pipelines using Vertex AI RAG Engine to enable accurate enterprise knowledge retrieval and context-aware responses.
  • Built AI workflow automation systems, reducing operational workload by 80% through intelligent process orchestration.
  • Developed LLM evaluation frameworks and observability dashboards, reducing hallucination rates and improving response quality monitoring.
May 2025 — Nov 2025

Sr Software Engineer — Gen AI

CSC
  • Developed RAG-powered document intelligence applications, automating enterprise document extraction workflows using Ollama, FastAPI, Docker, and Gradio.
  • Customized OpenWebUI with sensitive data masking, prompt sanitization, and privacy controls using SpaCy & NLP techniques.
  • Built AI-driven solutions for banking assessments, KYC classification, and compliance automation.
  • Developed an enterprise Document QA platform processing 5,000+ documents daily using OpenAI, LangChain, FAISS, and OCR pipelines.
  • Implemented dynamic prompt generation using user context, document metadata, and retrieval outputs to improve response accuracy.
  • Developed a Text-to-SQL analytics assistant using Vanna AI & MCP, enabling secure natural language interaction with enterprise databases.
Dec 2023 — May 2025

Machine Learning Engineer

Indian Tec Solutions
  • Built bank statement intelligence systems supporting 15+ banking formats, automating financial data extraction using OCR, Python, and custom document parsers.
  • Developed AI-driven KYC verification, fraud detection, and document processing solutions using Machine Learning, Computer Vision, and OCR.
  • Built predictive analytics and risk scoring models using TensorFlow and Python to support automated business decision-making.
  • Processed and analyzed 10TB+ of data using Python, Pandas, and MongoDB, enabling large-scale analytics and automation.
  • Developed AI-powered insurance claim and identity verification systems, reducing processing time from hours to minutes.
  • Deployed machine learning applications on AWS using Docker-based deployment pipelines.
June 2023 — Dec 2023

Data Scientist

LenDenClub
  • Developed machine learning solutions for credit risk assessment, customer analytics, and lending decision workflows using Python, Pandas, and MongoDB.
  • Built fraud detection and predictive risk scoring pipelines for financial documents using OpenCV, PyPDF, and Machine Learning techniques.
  • Developed AI-powered customer support automation systems, improving query resolution efficiency through intelligent workflows.
  • Designed scalable FastAPI-based services and contributed to codebase modularization through reusable internal Python packages.
  • Deployed ML applications on AWS using Docker-based deployment pipelines.
// selected work

Projects that moved real metrics.

Multi-Agent · ADK

Multi-Agent Customer Support & Automation Platform

Multi-agent AI assistant built with Gemini, Google ADK, Vertex AI RAG, guardrails and LLM evaluation — automating issue resolution, ticket creation and website assistance.

75% lower inference cost · 80% workload reduction
RAG · Document AI

Enterprise Document Intelligence & QA

Multi-modal RAG platform for document QA, OCR extraction, enterprise search and knowledge retrieval — OpenAI, LangChain, FAISS, Ollama, FastAPI, Gradio.

5,000+ documents processed daily
Text-to-SQL · MCP

Text-to-SQL Analytics Assistant

Natural-language to SQL analytics platform using Vanna AI and Model Context Protocol — secure NL access to enterprise databases with auto-generated reports, insights and charts.

Self-serve enterprise analytics
Privacy · NLP

OpenWebUI Privacy & Guardrails Layer

Customized OpenWebUI with sensitive data masking, prompt sanitization and privacy controls built on SpaCy and NLP — keeping enterprise LLM usage compliant.

PII-safe enterprise LLM access
KYC · Computer Vision

Smart KYC Verification & Insurance Automation

AI-powered identity verification, document validation and insurance claim automation using ML, Computer Vision and OCR — deployed on AWS.

Processing time from hours to minutes
FinTech · OCR

Financial Document Intelligence & Fraud Detection

OCR-driven bank statement analysis supporting 15+ banking formats with fraud detection and risk scoring pipelines using OpenCV, PyPDF and ML.

Stronger fraud signals across formats
RAG · Document AI

RAG Query Mail AI

End-to-end AI document processing & Q&A platform — PDF parsing with Surya OCR, PyMuPDF, PyTesseract, LlamaIndex and Qdrant vector DB. Email watcher auto-ingests documents into Qdrant.

Citation-aware responses powered by Ollama Llama 3.1
OCR · Identity AI

Intelligent KYC OCR

ML + OCR pipeline for document classification, Aadhaar masking and super-resolution APIs. TensorFlow, OpenCV, YOLO, NumPy — deployed on AWS.

97% reduction in document processing time
LLM · Chatbot

Document Q&A AI Chatbot

LangChain + GPT-4o Mini for precise retrieval over complex PDFs. FAISS vector store, OpenAI embeddings, PyMuPDF + PyTesseract OCR for scanned docs.

Enterprise document intelligence
Automation · NLP

Email Automation

Intelligent process automation for customer-support email categorization, failure log analysis and API monitoring with auto alerts.

15% decrease in errors
FinTech · ML

Banking Assessment Automation

Automated assessment of Banking Bureau Variables to predict loan repayment, with MongoDB-backed pipelines, XGBoost retention models and FastAPI services.

50% reduction in assessment time
ML · Risk

Fraud Detection

ML-based fraud detection catching tampered financial documents — advanced analytics and AI algorithms reducing fraudulent activity in bank statements.

Stronger transaction security
Vision · Research

Brain Tumor Detection

Published research paper in IJRASET (ISSN 2321-9653, ISRA Impact Factor 7.429) — applying ML & image processing for medical imaging.

Peer-reviewed publication
R&D · Alternate Data

Alternate Data Research

R&D for data enrichment and targeting — AI-driven solutions for customer profiling using internal and alternate data sources.

15% faster decision-making
// toolkit

The stack I ship with.

Agentic Systems

AI AgentsMulti-Agent SystemsGoogle ADKLangGraphLangFlowAgentic WorkflowsTool CallingMCP

Generative AI & RAG

LangChainLlamaIndexRAGVertex AI RAG EngineOpenAI GPT-4oGeminiOllamaPrompt EngineeringPrompt Caching

AI Safety & Eval

LLM EvaluationGuardrailsAI SafetyModel MonitoringOpenTelemetryAI ObservabilityFeature FlagsUnleash

Computer Vision & OCR

Surya OCRPyTesseractPyMuPDFOpenCVYOLOpdf2imageNLPSpaCy

Vector & Data Stores

QdrantFAISSChromaDBNeo4jPostgreSQLMongoDBCloud SQLRedis

ML Engineering

TensorFlowPyTorchXGBoostPandasNumPyText-to-SQLModel Deployment

Cloud & Infra

AWSBedrockGCPVertex AIAzureEC2S3Microsoft Copilot

Backend & APIs

PythonFastAPIREST APIsSSEDockerKubernetesGitCI/CD

AI Interfaces

StreamlitGradioReactNext.jsOpenWebUI
// recognition

Education, research & certifications.

Education

Bachelor of Engineering — Information Technology
Terna Engineering College · University of Mumbai
8.41 CGPA
Diploma in IT
Bharti Vidyapeeth Institute of Technology · M.S.B.T.E.
86.55%

Certifications & Achievements

  • Oracle Cloud Infrastructure 2024 Generative AI Certified Professional (87.5%)
  • Accenture Digital Skills — AI Certificate (90%)
  • Certified Accenture Developer Program
  • Machine Learning & Image Processing — NASSCOM
  • Microsoft Technology Associate — Python
  • IBM AI Foundation Badge
  • Published — Brain Tumor Detection · IJRASET (Impact Factor 7.429)
// let's build

Have an AI problem worth solving?

Always happy to talk about RAG, agents, OCR pipelines, and shipping AI products that actually run in production.