Prasad Chaudhari · Pune, India · Open to collaborations

Building Intelligent Products.Engineering Digital Experiences.

> Full Stack Developer

I'm Prasad Chaudhari, a Full Stack Developer and AI Product Builder focused on building scalable SaaS platforms, intelligent automation and AI-powered applications.

Explore My Products Let's Connect
  • 7+years building software
  • 5engineers I lead
  • 15+releases, 0 rollbacks
  • 100K+recipients per broadcast
  • 99.9%uptime
Download resume Ask my resume
trace · prasad.career()
runprasad.build(ai, scale)7+ yrs
├─ retrieveRAG · vector search+60% acc
├─ agentLangGraph · tools · memorymulti-step
├─ routeOpenRouter · GPT · Claude · Geminicost-aware
├─ fan-outSQS · Celery → 100K recipients2.4K/s
└─ deploy15+ releases0 rollbacks
status● ok · 99.9% uptime
Scroll
Live demo · RAG

Ask my resume.

A small version of what I build at work. Your question is matched against chunks of my resume, and the best matches go to a small LLM along with my full resume. It answers as me, with citations.

1 · Embed2 · Retrieve3 · Rerank4 · Generate
query embedding · 24 of 1,536 dims shown

Retrieval runs in your browser. Answers come from a small, low-cost LLM via OpenRouter, grounded only on my resume.

Knowledge base12 chunks · resume.md
Context window 0 / 3
About me

I turn AI ideas into systems people rely on.

I'm an AI Engineering Lead and Full Stack Architect with 7+ years of building cloud-native products. My focus is Generative AI: LangChain, RAG, LangGraph multi-agent workflows, and LangSmith for watching quality in production.

I also own the platform side. I set up a multi-tenant university admissions platform from the first client conversation to go-live: architecture, payments, messaging on four channels, and admin analytics.

I like the whole path from a rough requirement to a release that doesn't need a rollback.

Role
Engineering Team Lead
Company
Geta AI (SRV Media)
Based in
Pune, India
Team
5 engineers
Focus
GenAI · RAG · Agents
Education
B.E., PCCOE Pune
My journey

2015 to today, one span at a time.

Read it like an LLM trace: each row is a span, the bar shows how long it ran.

chapter
Toolbox

What I build with.

Selected work

Things I've shipped.

Featured · Multi-tenant SaaS

AI-Driven Application Form Platform

A university admissions platform I owned from client discussions and project setup to go-live. Several entrance exam programmes run on one backend, each with its own database.

  • Full applicant lifecycle: OTP signup, forms, documents, test and course selection, payments
  • BillDesk and Easebuzz payments with split settlement, a transaction ledger, and refunds
  • Messaging service for SMS (Smartping), WhatsApp, Email and Voice on Celery, SQS and Redis
  • Admin panel with role-based access, PII masking, 14+ reports and S3 exports
  • DigiLocker documents, AI photo checks, a RAG chatbot, and MongoMCP natural-language queries
3,000+TPS in load tests
100K+recipients per broadcast
~2.4K/smessage ingest
14+admin reports
100,000recipients reached
Illustration of one broadcast: an SQS queue feeds Celery workers that fan out to each channel.
Next.jsDjangoPythonMongoDB AtlasRedisCeleryAWS SQS · S3KubernetesBillDeskEasebuzzSmartpingWhatsApp APISentryk6
Generative AI

Geta AI Chatbot Platform

Multi-channel AI chatbot for WhatsApp and the web, built on RAG and LangGraph agents.

  • Cut manual support workload by 70%
  • Improved response accuracy by 60% with RAG
  • LangSmith tracing and evaluation in production
  • OpenRouter routing across OpenAI, Claude and Gemini
LangChainLangGraphLangSmithRAGOpenRouterNode.jsPython
Turtlemint

Schema-driven Form Builder

A React and JSON Schema framework reused across insurance products serving 5M+ users.

  • 60% less effort to build new product forms
  • Signzy KYC cut onboarding time by 40%
  • Advisor tools lifted lead-to-policy conversion by 30%
ReactTypeScriptJSON SchemaNode.js
Serverless

Billing & Notification Engine

Event-driven billing and alerts on Lambda, SQS and Stripe. Cut manual operations by 75%.

AWS LambdaSQSStripe
Messaging

Omnichannel Campaigns

WhatsApp, SMS and Email campaigns with queued delivery for 10K+ daily messages and live analytics.

PythonReactMongoDB
Data

CRM Contact Data System

Central CRM with validation, de-duplication and normalisation pipelines, plus secure lead-ingestion APIs.

PythonRESTAWS
How I work

From the first call to go-live.

The four steps I followed to set up the admissions platform. It plays on its own. Tap a step to jump to it.

client-call.md● live
Illustration based on how I delivered the admissions platform.
Impact

Results, measured.

70%

less manual support work with the AI chatbot

60%

better response accuracy with RAG

3,000+

TPS on the admissions platform under load

99.9%

uptime across projects

15+

production releases with zero rollbacks

5M+

users served by products I built at Turtlemint

45%

faster query resolution through CRM integrations

3×

recognised as a top performer

A normal day across my systems
Product studio

Products I'm building on my own.

Besides my work at Geta AI, I design and build my own SaaS products, from the first idea to the architecture and code.

See all products
Contact

Have an Idea? Let's Build Something Meaningful.

Hiring for engineering leadership, planning a SaaS product, or need help with RAG, agents or a platform that has to scale? I'm open to roles and collaborations. Send me a note.

Download my resume (PDF)
Email me LinkedIn GitHub Resume Save contact
Emailpnchaudhari1996@gmail.com
LinkedInlinkedin.com/in/prasad-chaudhari1996
GitHubgithub.com/prasad98765
LocationPune, Maharashtra, India
© 2026 Prasad Nimbaji ChaudhariBuilt with Next.js, FastAPI and a little canvas.
PrasadHi, I'm Prasad! Scroll down and I'll walk you through my work.
0/12