available for opportunities · bengaluru

Nagapoojith
Ullam
Builds AI that ships.

B.Tech AI/ML graduate (CGPA 8.2) who has designed and deployed 4 production-grade AI systems end to end — from multi-agent orchestrators on GCP to enterprise RAG platforms and full-stack SaaS products. AWS & Azure certified.

4
Production AI Systems
98%
ML Accuracy Achieved
120+
Teams Beaten — Hackathon
View Projects Get In Touch
AI
about me

Engineer. Builder. Shipper.

I'm a B.Tech AI/ML graduate from Chalapathi Institute of Technology (JNTUK, 2025) based in Bengaluru, with a CGPA of 8.2. I specialize in building production-grade Agentic AI systems, Python backends, and full-stack applications that actually get deployed.

In the past year, I've shipped 4 end-to-end AI systems — including a 3-node distributed multi-agent orchestrator running on GCP Vertex AI, an enterprise RAG platform, a multi-tenant SaaS with OAuth2/JWT security, and a clinical AI analyst achieving 98% accuracy.

I'm not a tutorial developer. Every project I build handles real engineering challenges: fixing GraphRecursionError infinite loops, designing A2A networking protocols, achieving zero-downtime deployments, and eliminating LLM hallucinations in production systems.

Currently open to full-time roles as Agentic AI Engineer, Python/Backend Developer, or Full Stack Developer in Bengaluru or remote.

Degree
B.Tech — AI & Machine Learning
University
JNTUK — Chalapathi Institute
CGPA
8.2 / 10 — Class of 2025
Certifications
AWS CLF-C02 · Azure AZ-900 · GCP GenAI
Location
Bengaluru, Karnataka, India
Status
● Open to opportunities
technical skills

Full-Stack Expertise

Agentic AI
LangGraphLangChainReAct AgentsMulti-Agent OrchestrationA2A NetworkingMCPCrewAILangSmithTool CallingMemory Systems
GenAI / LLMs
RAG PipelinesFAISSGemini 2.5 FlashOpenAI APIPrompt EngineeringHuggingFacePydantic V2GuardrailsVector Search
Python / Backend
Python 3.12FastAPISQLAlchemy 2.0Pydantic V2OAuth2 / JWTAlembicAsync/AwaitPytestDjangoREST API
Frontend
React.jsJavaScript ES6+HTML5 / CSS3BootstrapTailwind CSSjQuery / AjaxViteWordPressGatsby
Cloud / DevOps
GCP (Vertex AI · Cloud Run)DockerAWS (EC2 · S3 · Lambda)AzureGitHub Actions CI/CDDocker ComposeVercel · Render
Database / ML
PostgreSQLMongoDBScikit-learnXGBoost · Random ForestPandas · NumPyTensorFlow / KerasMLflowSageMaker
projects

Things I've Built & Shipped

Orchestrator LangGraph Coder Microservice MCP Sandbox GCP Vertex AI · Cloud Run
🤖 Agentic AI
Autonomous Software Engineer Orchestrator
3-node multi-agent AI system (Orchestrator → Coder Microservice → MCP Sandbox) using LangGraph state machines and A2A/ACP networking. Autonomously writes, executes, and debugs code with zero human intervention. Fixed critical GraphRecursionError by redesigning A2A payloads with 5-node execution history injection.
Zero infinite loops <300ms UI latency
LangGraphGemini 2.5 FlashFastAPIGCP Vertex AIReact.jsDocker
GitHub
Nexus AI Workspace Tenant 1 Tenant 2 Tenant N ReAct Agent ReAct Agent ReAct Agent OAuth2/JWT OAuth2/JWT OAuth2/JWT
⚡ Full Stack SaaS
Nexus AI Workspace — Multi-Tenant SaaS
Enterprise multi-tenant SaaS platform where users deploy specialized autonomous ReAct AI agents with long-term relational memory. Full React.js frontend + FastAPI backend with OAuth2/JWT auth, rate limiting (10 req/min/user), async LLM task offloading, and zero-downtime Alembic migrations.
100% env parity Zero downtime deploys
FastAPIReact.jsPostgreSQLSQLAlchemy 2.0Docker ComposeVercel/Render
GitHub
Supervisor Agent FAISS RAG Clinical PDFs Pandas Patient CSVs 98% accuracy ✓
🏥 AI / ML
Autonomous Healthcare AI Analyst
Dual-engine Supervisor agent that auto-routes between a FAISS RAG pipeline on clinical PDFs and a self-healing Pandas code interpreter on patient CSVs. Eliminated LLM quantitative hallucinations via System Persona injection and Guardrails validation into LangGraph state.
98% accuracy 60% latency reduction
LangGraphFAISSHuggingFaceGemini 2.5 FlashFastAPIReact.js
GitHub
Documents FAISS Vector Index FastAPI +70% precision HuggingFace Embeddings · PostgreSQL
🔍 RAG / Backend
Enterprise RAG System
Production-grade Retrieval-Augmented Generation platform with FAISS vector indexing, HuggingFace embeddings, and async FastAPI endpoints. Built comprehensive Pytest test suites covering 100% of API routes with zero regression failures across 3 deployment iterations.
70% better retrieval Zero regressions
FastAPILangChainFAISSHuggingFacePostgreSQLDocker
GitHub
Network Traffic Analysis Benign ✓ Botnet ✗ Random Forest + XGBoost 98% acc. Django Web App CSV Upload + Predict Final Year Project · Team Lead · 4 members
🔬 ML / Team Lead
IoT Botnet Detection — Two-Fold ML System
Led a team of 4 to build a real-time IoT network traffic classifier. Owned the complete ML lifecycle: data collection, feature engineering on packet counts, training Random Forest + XGBoost ensemble, and deploying via joblib into a Django MVT web app with visual distribution charts.
98% accuracy Team of 4
PythonScikit-learnXGBoostDjangoMySQLMatplotlib
GitHub
FastAPI Gateway Pydantic V2 Gemini Pro OpenAI SDK React.js Frontend Vercel GitHub Actions CI/CD · Zero manual deploys
🔗 Full Stack
Nexus LLM Gateway — Unified AI API
Cloud-native LLM gateway unifying Gemini and OpenAI behind a single FastAPI endpoint with Pydantic V2 structured output validation, server-side API key security, CORS enforcement, and async non-blocking processing. Automated GitHub Actions CI/CD with zero manual deployment steps.
Zero manual deploys Multi-LLM support
FastAPIReact.jsGemini ProOpenAI SDKVercelGitHub Actions
GitHub
🏆

Hackathon Winner — 1st Place

Secured 1st place out of 120+ competing teams — designed and shipped a complete, working AI solution within a 24-hour build window under pressure.

experience

Work Timeline

Jan 2025 — Present
Agentic AI Engineer — Independent Projects
Self-Employed · Bengaluru, India · Remote
  • Architected a 3-node multi-agent system on GCP Vertex AI using LangGraph state machines and A2A/ACP networking — fully autonomous code research, generation, sandboxed execution, and self-debugging with zero human intervention.
  • Engineered a multi-tenant SaaS platform with React.js + FastAPI, OAuth2/JWT auth, rate limiting, async LLM task offloading, and 100% dev-to-production parity via Docker Compose.
  • Built a dual-engine Healthcare AI Analyst achieving 98% accuracy on patient metric calculations by eliminating LLM hallucinations through System Persona injection and Guardrails validation.
  • Shipped 4 production-grade AI systems, all containerized with Docker and deployed on GCP/AWS/Vercel/Render with automated CI/CD pipelines.
LangGraphFastAPIGCPReact.jsDockerPostgreSQL
Dec 2024 — Mar 2025
AI/ML Intern
Corporate Innovative Technologies Pvt. Ltd. · AICTE & APSCHE Recognized
  • Applied supervised learning techniques including classification and regression using Python, Scikit-learn, and Pandas on real-world datasets across a 4-month program.
  • Built end-to-end ML pipelines covering data preprocessing, feature engineering, model training, and evaluation using precision, recall, F1-score metrics.
  • Gained hands-on experience with NumPy and Matplotlib for data analysis and visualization of model performance metrics.
PythonScikit-learnPandasNumPyMatplotlib
Aug 2024 — Apr 2025
Final Year Project Lead — AI/ML
Chalapathi Institute of Technology, JNTUK · Team of 4
  • Led a team of 4 to build a two-fold ML system classifying IoT network traffic as benign or botnet in real time, achieving 98% accuracy on a large-scale network dataset.
  • Owned the full ML lifecycle: data collection, feature engineering on packet count patterns, training Random Forest and XGBoost ensemble models, evaluation using precision/recall/F1.
  • Deployed the best model via joblib into a Django MVT web app — users upload CSV files and receive real-time predictions with visual distribution charts.
PythonScikit-learnXGBoostDjangoMySQL
education & certifications

Academic Background

B.TECH · 2021–2025
Artificial Intelligence & Machine Learning
Chalapathi Institute of Technology, JNTUK
Guntur, Andhra Pradesh · CGPA: 8.2 / 10
CERTIFICATIONS
☁️
AWS Cloud Practitioner (CLF-C02)
Amazon Web Services · 2024
🔷
Azure Fundamentals (AZ-900)
Microsoft Azure · 2024
🤖
GCP Generative AI Fundamentals
Google Cloud · 2025
contact

Let's Build Together

Open to full-time roles in Bengaluru or remote.

Looking for Agentic AI Engineer, Python/Backend Developer, or Full Stack Developer positions. If you're building something ambitious and need someone who can ship production AI systems from day one — let's talk.

email
ullamnagapoojith222@gmail.com
linkedin
nagapoojith-ullam
github
github.com/Unp12
location
Bengaluru, Karnataka, India
// send_message()