Building agentic AI systems

Naveen Kunisetty

Senior Staff Developer at Oracle (previously Architect at Tractor Supply Co.) with 15+ years shipping enterprise software. Today I design RAG systems, AI agents and LLM-powered products on scalable cloud-native platforms.

// About

From mobile apps to intelligent systems

I started as a Java and iOS engineer, grew into a full stack lead and then an architect at Tractor Supply Co. Today I build AI-first platforms as a Senior Staff Developer at Oracle. I care about the whole path: retrieval quality, agent reliability, latency, cost, security and the people who use it.

0Years in software
0Companies, 4 industries
0GPA in M.S. Computer Science
0AI / LLM solutions shipped
// AI Work

What I build with AI

Production AI, not demos: grounded answers, agents that can be trusted with tools, and the evaluation and guardrails around them.

RAGAgenticLLM appsVoice & NLPCloud
Retrieval

RAG Systems

Retrieval-augmented generation over enterprise knowledge: chunking, embeddings, hybrid search, re-ranking and citation-grounded answers.

Vector DBsEmbeddingsAzure AI SearchRe-ranking
Agentic

AI Agents

Tool-using agents with planning, memory and human-in-the-loop checks. Multi-agent workflows with clear boundaries and audit trails.

ClaudeOpenAIMCPLangGraph
Generative

LLM Applications

Azure OpenAI GPT copilots and assistants, prompt engineering, structured outputs, evals and token-cost optimization.

Azure OpenAIPromptingEvalsGuardrails
Speech

Voice & NLP

Wake-word detection, voice biometrics and natural language processing for hands-free, secure experiences.

Wake WordVoice IDNLP
Platform

Cloud & Platform

Cloud-native microservices on Azure and Cloudflare with CI/CD, fault tolerance, security audits and observability.

AKSService BusWorkersGitHub Actions

New knowledge loading...

$ learning agent evals

Always learning. Currently exploring what comes next in agentic AI, and updating this page as I build it.

How a RAG answer is made
Queryuser asks Embedvectors Retrievehybrid search Re-ranktop-k context LLM+ guardrails Answerwith citations
// Projects

Featured builds

Live · Agentic AI

Trading Agent

An autonomous trading agent that scans the market daily, reasons over a fact sheet with an LLM, and places paper trades, with a public dashboard of every decision.

  • Ports-and-adapters design: swap model, vendor or agent SDK by config
  • Rule engine + LLM judgment, with token and cost tracking per decision
  • Real-time Discord alerts for every trade, the daily recap and system health
Open live dashboard →
Live · Multi-Agents Game

Teen Patti Agents

Take a seat at a Teen Patti table against an AI player, while an AI dealer runs the game. Enter your name, look at your cards or play blind, then chaal, raise, show or pack.

  • Multi-agent turn-taking, with a persona for every seat
  • Deterministic rules engine; the LLMs only choose legal moves and talk
  • You play live: only legal moves are offered, and every hand ends with a winner reveal
Deal me in →
dealer  Raju: Boot of 2 collected. Cards dealt!
meera   see       "Let me check the numbers."
meera   raise 8   "I like what I'm holding."
you     see   → K♥ K♠ 7♦ (Pair)
you     raise 16
meera   chaal 16  "I'm not scared of you."
you     show
result  You win 52 · Pair of K beats Pair of 6
next?   [Deal hand 2]  [End game]
// Experience

Where I've shipped

Mar 2026 — Present

Senior Staff Developer

Oracle

Build AI-first platforms: RAG systems, AI agents and LLM-powered applications at enterprise scale.

2018 — 2026

Software IT Architect · Full Stack Lead

Tractor Supply Co · Brentwood, TN

Architected cloud-native and AI platforms: RAG assistants, LLM agents on Azure OpenAI, voice and vision features in mobile apps. Led cross-functional teams, defined architecture blueprints and mentored engineers.

2017 — 2018

Senior iOS Developer

BBVA Compass · Dallas, TX

Led development of the Tuyyo money-transfer app with Swift and Java.

2016 — 2017

Software Developer

Citrix · Santa Clara, CA

Built the WorxHome enterprise mobility app for iPhone and iPad.

2012 — 2014

Senior Member Technical

ADP · Hyderabad

iOS developer on the ADP mobile payroll app.

2010 — 2012

Associate Consultant

Capgemini · Hyderabad

Java APIs for a financial services application.

Education

M.S. Computer Science · GPA 4.0

Northwest Missouri State University · B.E. Computer Science, Osmania University
// Toolbox

Stack

// Contact

Let's build something intelligent

Open to conversations about AI architecture, agents and RAG.