LogoHemanth Puppala

Hemanth Puppala

AI/ML Engineer

AI and Software Engineer with 3+ years of experience building and deploying LLM-powered agent systems, RAG pipelines, and production AI infrastructure. I ship reliable, scalable systems — from fine-tuning models to orchestrating multi-agent workflows in enterprise environments.

0+
Years Experience
0+
Projects
0+
Daily Transactions
Scroll

About Me

I'm an AI and Software Engineer with 3+ years of experience building and deploying LLM-powered agent systems, RAG pipelines, and production AI infrastructure. Proficient in LangGraph, LangChain, and agentic workflow orchestration with hands-on experience integrating AI into enterprise business workflows via REST APIs.

My journey started with computer vision — building real-time object detection, autonomous navigation, and edge-deployed models on Jetson hardware. I then moved into generative AI, designing stateful multi-agent systems with tool calling, function orchestration, and persistent conversation memory.

Most recently, I've been building LLM-powered enterprise systems at Denali Advanced Integration — designing end-to-end agentic pipelines for order validation, PO/SO generation, and inventory tracking. I also co-founded MendIt, where I architected and shipped a full-stack platform as the sole engineer across 12 modules.

LangGraphLangChainLlamaIndexRAGFine-TuningPyTorchTensorFlowComputer VisionCUDAFAISSDockerAWSFastAPINode.jsTypeScript
Hemanth Puppala

Projects

Demo slide 1
Demo slide 2
Demo slide 3
Demo slide 4
Demo slide 5
Demo slide 6
Demo slide 7
Demo slide 8
Demo slide 9
2025

Ekam CLI - VLM/LLM Testing Tool

Production-grade CLI unifying HuggingFace, Ollama, LlamaCPP, and vLLM for AI model testing, benchmarking, quantization, and fine-tuning. Used at Denali — reduced inference latency by 40-60% and scaled to 5,000+ stations.

PythonPyTorchvLLMOllamaLlamaCPPTypeScript

View details →

Cognitive Research Hub
AI-powered Research Engine
|
📎
Send
📄
deer_detection_paper.pdf

Upload document & ask question

2025

DocuMentor - AI Research Platform

Upload a document with a research question — 8 LangGraph agents orchestrate parallel web and arXiv searches, synthesize a cited report, support two follow-up modes, and generate AI-graded quizzes.

FastAPILangGraphLangChainGeminiGroqBrave API

View details →

Context-Aware Agentic AI System
2025

Context-Aware Agentic AI System

YAML-driven multi-agent framework powering this chatbot — config-driven tool construction, category-filtered RAG on Pinecone/Cohere, streaming orchestration with fallback execution, and input/output guardrails.

LangGraphLangChainPineconeCohereFastAPIOpenRouter

View details →

What I'm Building

Ventures and products currently in development

Live

MendIt - On-Demand Repair Marketplace

An end-to-end platform connecting customers with nearby repair shops for electronics repairs. Built as sole engineer across 12 modules — scalable NestJS backend, PostgreSQL with PostGIS for geolocation, Redis caching, WebSocket/SSE real-time pipelines, Stripe integration, and 15+ RESTful API endpoints.

Node.jsNestJSPostgreSQLPostGISRedisStripeDocker
FounderSole Engineer
View Live Preview↗
In Development

SymLink - Self-Service Utility Platform

An infrastructure platform enabling self-service utility dispensing across India — water, EV charging, and fuel — through a unified consumer app and smart IoT hardware ecosystem.

FlutterDartESP32IoT SensorsDocker
Co-FounderPlatform Architect

Experience