Currently at Cisco · Building AI Systems

Bargavi
Kongara

Gen AI Engineer

I build intelligent systems that work in production — RAG pipelines, MCP servers, and LLM-powered agents that cut retrieval time by 40% and hit 99%+ accuracy.

Bargavi Kongara
scroll

From Quality to
Intelligence

I started my career ensuring software worked flawlessly — 13+ years of relentless attention to systems, data, and outcomes. Now I build the systems themselves. Today at Cisco, I design and ship RAG pipelines, MCP servers, and LLM-powered agents that move the needle.

My MS in Business Analytics from Clark University (GPA 3.96) gave me the analytical foundation; years in production gave me the instincts. I was inducted into Beta Gamma Sigma — the top 20% of business school graduates worldwide — for academic excellence.

I build things that are accurate, fast, and explainable. Not demos. Production-grade AI systems that engineers and stakeholders can trust.

🏛️

Education

MS Business Analytics, Clark University

Academic Standing

GPA 3.96 · Beta Gamma Sigma

🏢

Current Role

AI Engineer / Test Lead at Cisco

📍

Open To

Gen AI / MLRoles

What I Build

End-to-end AI systems — from retrieval architecture to deployed models.

🔍

RAG Pipelines

Production retrieval-augmented generation systems with optimized chunking, embeddings, and hybrid vector search.

⚙️

MCP Servers

Model Context Protocol servers that expose structured tools and enterprise data to LLM agents at scale.

🤖

AI Agents

Autonomous LLM-powered agents with tool use, memory, and multi-step reasoning for complex enterprise workflows.

📊

ML Models

Supervised and ensemble models for fraud detection and prediction — built to achieve 99%+ accuracy in production.

Experience

13+ years across enterprise tech, building systems at every layer of the stack.

Cisco Systems

AI Engineer / Test Lead · San Jose, CA

Apr 2024 — Present

Built and deployed a RAG pipeline using LangChain + FAISS, indexing 50+ documents — reducing engineer retrieval time by 40%. Optimized RAG retrieval precision by 30% through custom chunking, metadata filtering, and embedding tuning. Designed and shipped 3 MCP servers in Python, enabling LLM agents to autonomously call internal Cisco APIs and tools, saving ~5 hrs/week across the team. Applied prompt engineering and agentic AI patterns to automate test scenario generation and intelligent test prioritization. Led enterprise E2E QA delivery across 20+ cross-functional teams, achieving 100% cross-system data reconciliation via SQL.

Sri Infotech, Inc.

Data Analyst · Boston, MA

May 2023 — Dec 2023

Cleaned and analyzed 50k+ record datasets using Python and SQL. Built automated Power BI dashboards delivering actionable insights to business stakeholders.

Infosys (Client: Cisco)

Technical Test Lead · Hyderabad, India

Jan 2021 — Aug 2022

Automated 70% of regression tests using Java, Selenium WebDriver (POM), TestNG, and Maven — significantly accelerating release cycles for an enterprise AngularJS application. Reduced defect leakage by 30% and defect turnaround time by 20% through structured testing. Delivered executive-level QA reporting to senior stakeholders.

IGT Pvt Ltd

SQA Engineer II · Hyderabad, India

Jun 2016 — Dec 2020

Performed functional, performance, and API testing using Postman and Rest Assured. Identified 50+ critical defects, reducing post-release issues by 20% and improving test reliability by 25%.

ADP India Pvt Ltd

Member Technical · Hyderabad, India

Jun 2011 — May 2016

Delivered quality validation across PeopleSoft HCM modules. Authored 300+ test cases and introduced Java/Selenium automation, reducing manual testing effort by 30%.

Featured Projects

Production systems, not side projects. Built to measure.

🔍 RAG / Retrieval

Enterprise RAG Pipeline at Cisco

Built and deployed a RAG pipeline using LangChain + FAISS, indexing 50+ internal Cisco documents. Replaced manual knowledge-based searches with LLM-grounded responses via OpenAI API. Further optimized retrieval precision by 30% through custom chunking strategies and embedding tuning.

⚡ 40% reduction in retrieval time · 30% precision gain
LangChainFAISSChromaDBOpenAI APIPython
⚙️ MCP / Tooling

MCP Server Implementation

Built 3 production MCP servers exposing enterprise data sources, APIs, and compute resources as structured tools for LLM agents. Enables autonomous workflows across Cisco's AI platform infrastructure.

🚀 3 servers in production
MCP ProtocolPythonFastAPIDocker
🔒 ML / Fraud

Supply Chain Fraud Detection

Built ML models to predict fraud and late delivery in supply chain data. Used Logistic Regression, Random Forest, and Decision Tree with PySpark + scikit-learn across a large-scale dataset. Performed end-to-end feature engineering and model evaluation.

🎯 99.12% model accuracy
PySparkScikit-learnLogistic RegressionRandom ForestPython
💰 ML / Finance

Financial Fraud Prediction

Applied logistic regression to detect fraudulent payment transactions. Performed end-to-end data exploration, cleaning, and model evaluation on financial services data to minimize false positives while maintaining near-perfect recall.

🎯 99.91% model accuracy
Logistic RegressionScikit-learnPandasPythonSQL

Skills

Technologies I use to build intelligent systems end-to-end.

Gen AI & LLM
LangChainLlamaIndexOpenAI API Anthropic Claude APIRAGMCP Servers AI AgentsPrompt Engineering Tool-CallingEmbeddings FAISSChromaDB
ML & Data
PythonPandasNumPy Scikit-learnPySpark Logistic RegressionRandom Forest Decision TreePower BISQL
Cloud & DevOps
AWSMicrosoft AzureGCP JenkinsGitGitHub BitbucketCI/CD
Programming
PythonJava JavaScriptTypeScript
Other
REST APIsPostmanJIRA ConfluenceAgile / Scrum

Education & Certifications

Graduate Degree

Clark University

Master of Science in Business Analytics

⭐ GPA 3.96🏅 Beta Gamma Sigma

Certification

Python Institute

PCEP — Certified Entry-Level Python Programmer

✅ Certified

Certification

Google

Google Analytics Certification

✅ Certified

Let's Build
Something Real

Open to Gen AI/ML engineering roles, research collaborations, and advisory conversations. If you're working on something serious, let's talk.