GENERATIVE AI ENGINEER

Vamshi
Yelagandula
.

3 years of experience building end-to-end GenAI solutions — intelligent chatbots, RAG pipelines, LangGraph multi-agent systems, and document AI that solves real business problems at scale.

View My Work Download Resume ↓
3+
years in GenAI
2
product companies
4+
AI projects shipped
55%
processing time saved
ABOUT ME

Turning prompts
into products.

I'm Vamshi Yelagandula, a Generative AI Engineer with 3 years of experience specialising in Large Language Models, RAG architectures, and production-grade AI applications.

I build end-to-end GenAI solutions — from intelligent enterprise conversational agents to autonomous multi-agent research systems — using LangChain, LangGraph, OpenAI, Hugging Face, and FastAPI, deployed on AWS.

Experienced across enterprise CX automation, decision intelligence, and document AI domains. Passionate about translating complex AI capabilities into scalable, cloud-hosted, business-ready products.

LOCATION
Bangalore, India 🇮🇳
FOCUS
LLMs · RAG · AI Agents
STATUS
● Open to Opportunities
~$ cat profile.json
"name": "Vamshi Yelagandula",
"role": "Generative AI Engineer",
"company": "Fractal Analytics",
"years_exp": 3,
"specialties": [
"RAG Pipelines",
"Multi-Agent Systems",
"LangGraph",
"LLM Applications"
],
"open_to_work": true
~$ ▌
SKILLS & TOOLS

My tech stack

Technologies I use to design, build, and ship production-grade GenAI solutions end-to-end.

🤖
Generative AI & LLMs
LangChainLangGraph OpenAI APIHugging Face TransformersRAG Pipelines Prompt EngineeringAI Agents Multi-Agent SystemsLLM Observability Fine-Tuning
🛠️
Frameworks & Tools
FastAPIStreamlit LangSmithGroq SDK Tavily SearchBeautifulSoup
🗄️
Vector Databases & Retrieval
FAISSChromaDB Embedding Pipelines Similarity Search
💻
Programming & Data
PythonPandas SQLJSON/CSV Processing Data WranglingREST APIs
☁️
Cloud & Deployment
AWS EC2AWS S3 AWS IAMGit GitHubCI/CD
SELECTED WORK

Projects that ship.

Production AI applications — deployed, live, and accessible to users right now.

01
Multi-Agent AI Research System
Autonomous research pipeline built with LangChain's create_react_agent, coordinating a Search Agent (Tavily API) and a Reader Agent (BeautifulSoup) for automated topic research and web scraping. A Writer Chain (LangChain LCEL + Groq LLaMA 3.3 70B) synthesises findings into a structured report, and a Critic Chain provides automated quality scoring out of 10 with a final verdict.
LangChainLangGraph Groq LLaMA 3.3 70BTavily Search BeautifulSoupMulti-Agent
02 LIVE
InvoiceIQ — Smart Invoice Data Extractor
Streamlit web app that takes GST invoice PDFs and extracts fully structured data with zero manual effort. Uses PyMuPDF to rasterise PDF pages into images, feeds them into Groq Llama 4 Vision with strict validation rules for IRN, GSTIN, date formats, and numeric fields. Outputs organised into logical sections with JSON and CSV download options.
PythonStreamlit Groq SDKLlama 4 Vision PyMuPDFPrompt Eng.
Live Demo →
03 LIVE
ICICI Bank Customer Support Assistant
RAG-based banking chatbot with a full LangChain pipeline — WebBaseLoader, RecursiveCharacterTextSplitter, HuggingFace embeddings, and FAISS vector store. LangChain agent routes queries through a rag_tool with intelligent fallback to Tavily web search for real-time RBI updates. Includes domain validation, greeting detection, and multi-turn chat history.
LangChainFAISS HuggingFaceStreamlit Tavily SearchRAG
Live Demo →
04
AI Agent Dashboard — Multi-Tool LLM Assistant
Multi-functional AI agent capable of handling weather queries, note-taking, and task management in a unified Streamlit chat interface. Custom tool-calling architecture (get_weather, save_note, add_task) built with LangChain and Groq LLaMA 3.3 70B. Features persistent JSON-based CRUD storage, session memory, conversation history, and robust error handling.
LangChainGroq LLaMA 3.3 70BStreamlit AI AgentsTool Calling
WORK HISTORY

Where I've worked.

Building production AI at Fractal Analytics and Yellow.ai — 3 years of real-world GenAI engineering.

JAN 2025 — PRESENT
Generative AI Engineer
FRACTAL ANALYTICS · Bangalore
  • Developed and deployed enterprise-grade LLM-powered decision intelligence pipelines using LangChain and LangGraph, enabling multi-step autonomous reasoning — accelerating report generation cycles by ~40%.
  • Built Multi-Agent AI systems coordinating specialised agents (research, retrieval, synthesis) using LangGraph state machines to automate complex enterprise data workflows end-to-end, reducing workflow completion time by ~35%.
  • Engineered production RAG pipelines with FAISS and HuggingFace embeddings for enterprise knowledge bases, optimising chunking strategies — improving retrieval precision by ~30%.
  • Developed AI model pipelines as scalable REST microservices using FastAPI, supporting 500+ daily API calls with <200 ms average response time.
  • Implemented end-to-end LLM observability using LangSmith — tracing prompts, monitoring token usage, and debugging chains — reducing production error rates by ~25%.
  • Deployed Streamlit and FastAPI apps to AWS EC2, managed static assets via AWS S3, and configured IAM roles and security groups for secure cloud-hosted AI delivery.
  • Integrated multimodal AI capabilities (vision models, document AI) to automate structured data extraction from PDFs, invoices, and scanned enterprise documents.
  • Collaborated with data science and product teams to translate business requirements into reliable GenAI features — reducing manual data processing time by ~55%.
  • Implemented prompt engineering workflows for OpenAI GPT and open-source LLMs (LLaMA, Mistral) to improve response accuracy and consistency across enterprise use cases.
JUN 2023 — DEC 2024
Generative AI Engineer
YELLOW.AI · Bangalore
  • Designed and deployed conversational AI agents for enterprise CX automation using LangChain — handling 10,000+ customer queries monthly with an ~85% self-service containment rate.
  • Built RAG-based Conversational Knowledge Base pipelines using WebBaseLoader, RecursiveCharacterTextSplitter, HuggingFace embeddings, and FAISS — improving answer relevance scores by ~40% over keyword-based baselines.
  • Developed tool-orchestrated LLM agents with intelligent query routing — primary RAG retrieval with real-time Tavily web search fallback — ensuring high-coverage responses across diverse customer intents.
  • Implemented domain validation, greeting detection, and multi-turn chat history within conversational agents for contextually relevant and secure customer interactions.
  • Engineered and fine-tuned prompts for enterprise-grade customer support LLMs, incorporating tone consistency, safety guardrails, and structured output formatting.
  • Integrated vector search workflows (FAISS, ChromaDB) into chatbot backends — reducing average query latency by ~30% across BFSI and retail domains.
  • Leveraged LangSmith for end-to-end LLM tracing and observability across all conversational agent pipelines — cutting average debugging time by ~50%.
  • Developed REST APIs using FastAPI to expose conversational AI pipelines as microservices, enabling seamless integration with enterprise CRM and helpdesk platforms.
2017 — 2021
B.Tech — Mechanical Engineering
ANURAG GROUP OF INSTITUTIONS · Hyderabad
GET IN TOUCH

Let's build
something great.

Open to full-time GenAI roles, freelance AI projects, and consulting. Whether it's a RAG system, a LangGraph multi-agent pipeline, or a document intelligence tool — let's talk.

vamshiyelagandula15@gmail.com +91 9281292718
Download Resume ↓