Available for AI Engineer roles · OPT · July 2026

Building AI products
that actually ship.

AI Engineer at Scale AI. Founder of UNFROZEN — a live AI coaching SaaS with real paying users and ~97% gross margins. MS in Data & AI from UNT. I turn ideas into production systems.

UNFROZEN · Production Metrics
~97%
Gross Margin
$0.09/user
Inference Cost
<2s
AI Response
$15/mo
Infrastructure
Hemanth Anuginti
Hemanth Anuginti
AI Engineer · Denton, TX, USA
Scale AI · Active
🚀Flagship Product

UNFROZEN EMBER

AI Coaching Platform · Live with real paying users

⚡ UNFROZEN
LIVE
EMBER Memory Engine
"Based on our previous 12 sessions, I can see you've been working on vocal clarity — let's build on that today with targeted exercises..."
GPT-4o-mini
Whisper
Stripe
Supabase
Next.js 16
FastAPI
✓ Real Paying Users✓ Solo Founder✓ ~97% Gross Margin

Built end-to-end solo — from system design through production deployment and incident response. EMBER is a custom cross-session memory architecture using context injection instead of RAG: a single PostgreSQL query injects full user history into the prompt, eliminating retrieval latency while guaranteeing perfect-fidelity recall, returning structured JSON evaluations in under 2 seconds.

What I Built
•Cross-session EMBER memory engine
•Stripe payment + signature-verified webhooks
•OpenAI Whisper speech transcription
•Row-Level Security (Supabase)
Tech Stack
•Next.js 16 + TypeScript
•Python FastAPI backend
•Supabase (PostgreSQL)
•Vercel + Render CI/CD
Visit unfrozen.me

Live Projects Portfolio

All deployed and ready to explore

UNFROZENLive · Paying Users

AI Coaching Platform

Production AI coaching SaaS featuring EMBER — a custom cross-session memory engine using context injection instead of RAG. Ships at ~97% gross margins with real paying users. Built solo, end-to-end.

✓ ~97% gross margin
✓ $0.09/user inference
✓ Sub-2s AI response
✓ Real paying users
Next.js 16FastAPIGPT-4o-miniSupabaseStripeWhisper
RefactorAILive Demo

AI-Powered Code Analysis Engine

Java code refactoring engine using JavaParser AST + Groq LLM API. 7 code-smell detectors at 95% accuracy, JWT-secured REST APIs handling 100+ concurrent requests at sub-500ms.

✓ 95% detection accuracy
✓ 100+ concurrent req
✓ Sub-500ms response
✓ Monaco code editor
Spring BootReactJavaParser ASTGroq LLMDockerJWT
C
C
Churn PredictionAWS Production

Cloud ML Pipeline · AWS

End-to-end cloud-native ML pipeline on AWS processing 500K+ customer records. 40% runtime reduction via distributed Hive optimization. 79.9% model accuracy with Tableau dashboards.

✓ 500K+ records
✓ 79.9% accuracy
✓ 40% faster runtime
✓ Tableau dashboards
PySpark MLlibAWS S3/EMRHiveLambdaTableau
Academic Project

Where I've Worked

S
AI Engineer Intern
Scale AI
USA · Remote
Feb 2026 – Present
Current
→Agentic AI solutions with LangGraph & FastAPI
→MCP connectors for enterprise knowledge bases
→RAG pipelines using PostgreSQL + pgvector
→Multi-agent workflow patterns (planner/executor/reviewer)
→Prompt engineering & evaluation frameworks
C
Software Engineer
Chargebee
India
Jul 2021 – Jul 2024
→Scaled billing APIs processing thousands of daily transactions (30% perf gain)
→Microservices architecture — 26% reduction in dev effort
→PostgreSQL / MySQL query optimization (35% efficiency gain)
→CI/CD with Jenkins + GitHub Actions (32% fewer failures)
→Automated testing — 85%+ code coverage, 27% fewer incidents

Technical Expertise

OpenAI API (GPT-4o, Whisper)
LangChain / LangGraph
MCP (Model Context Protocol)
RAG Pipelines
Agentic AI Workflows
Context-Injection Memory (EMBER)
Prompt Engineering & Evaluation
Vector Databases (Pinecone, FAISS, ChromaDB)
AI Agents & Orchestration
Groq / Claude API
🤖
AI Fluency: Framework & Foundations
Anthropic · 2026
⚡
Claude 101
Anthropic · 2026
🧠
IBM ML Professional Certificate
IBM · 2025