Accepting consulting calls

Your AI project needs an architect, not another demo

I am S. Abhijit Rao, a freelance AI engineer. I have shipped production RAG systems, agentic chatbots, and OCR pipelines for energy and petroleum companies — and I now take a limited number of consulting calls to tell you what to build, what to cut, and what it will cost.

Book a free 30-min callSee pricing

No pitch deck. No sales sequence. Just an honest answer.

Scroll

Book a call

Find out if AI is right for your problem — in 30 minutes

No pitch deck, no discovery invoice. Tell me what you are trying to build and I will tell you whether it is worth building.

Book a free callWhatsApp me
Calls booked this monthLimited
Response timeWithin 24 hours
TimezoneIST — flexible worldwide
Emailsar.abhijit2003@gmail.comPhone+91 9985727779WhatsAppFastest reply

What I do

AI consulting for teams who need it to actually work

Three ways I work with founders and engineering teams. Most engagements start with the first one.

Architecture

AI Architecture Review

Leave a 90-minute call with a written, buildable plan — not a slide deck of what-if.

$50

Who it is for: Founders and eng leads who have an AI idea, an existing prototype, or a chatbot that underperforms, and need someone to tell them what to actually build.

  • Live teardown of your current stack, data, and bottlenecks
  • A written architecture recommendation you can hand to any engineer
  • Cost and timeline estimate for the first shippable version
  • An honest verdict on whether AI is the right tool at all
RAG & Agents

RAG & Agent Troubleshooting

For chatbots that hallucinate, ignore your documents, or answer the same question three different ways.

$50

Who it is for: Teams who already shipped a retrieval-based assistant and are now living with the consequences.

  • Diagnosis of chunking, retrieval, and prompt failure points
  • Fix recommendations ranked by effort vs. impact
  • Evaluation strategy so you can measure the improvement
Production

Taking AI to Production

The gap between a working notebook and software you can charge money for is where most AI projects die.

$50

Who it is for: Teams with a prototype and no path to reliability, cost control, or monitoring.

  • Deployment, cost, and reliability review
  • Model and vendor strategy — including when to stop using an LLM
  • A realistic build-vs-buy recommendation

Pricing

Published rates, because guessing wastes both our time

Every engagement is fixed-price and agreed in writing before any work starts.

Clarity Call

30 min · video

Free

No pitch, no obligation

You have an idea and want a straight answer on whether it is worth building.

  • 30 minutes on a video call
  • Your problem explained and pressure-tested
  • A clear recommendation on what to do next
Book a free call
Most booked

Architecture Deep-Dive

90 min · video + written plan

$50

Pay before the call

You need a real plan — a working team, a budget, or a board asking questions you cannot answer.

  • 90 minutes of deep technical discussion
  • Teardown of your current or proposed stack
  • A written architecture plan, delivered within 48 hours
  • Cost and timeline estimate
  • Two follow-up email questions for 14 days
Book the deep-dive

Build Partnership

Ongoing · scoped projects

From $500

Scoped after the call

You liked the plan and want it actually built, by the person who wrote it.

  • Fixed scope and price, agreed in writing first
  • Direct build work — no handoff, no account manager
  • Weekly demos against a written spec
  • You own the code and the infrastructure
Discuss a project

All prices in USD. Need something else? Ask about a custom scope.

Proof

I have done this before, for paying clients

Not a course. Not a side project. Production systems shipped for companies, several of them still running.

Nuevosol Energy

Engineers were spending hours manually searching internal technical documents to answer routine questions.

Built the Agentic RAG solution

Cut manual file-search time with a retrieval chatbot built on LangGraph and ChromaDB.

Prodigal AI

A video-editing platform needed to turn raw footage into publish-ready content without manual timeline work.

Agentic AI Intern

Shipped automation on the Dhanur AI platform and was named Intern of the Month, April 2025.

Astraveda

Petroleum analytics needed both a conversational assistant over data and OCR on scanned documents.

Freelance AI Engineer — full ownership

Architected and shipped Ask Astra (chatbot) and Click Astra (OCR) in production, end to end: Next.js, FastAPI, LangGraph, AWS, Supabase.

Visit petro-astra.in

How it works

From first message to written plan in three days

You will know exactly what happens after you click book. No surprises, no mystery invoices.

  1. 01

    You book a slot

    Pick a time that works. A short intake note comes with the invite so I can read your situation before we speak.

  2. 02

    We talk for 30 minutes

    You describe the problem, I ask the uncomfortable questions. You will leave this call knowing whether to build, buy, or stop.

  3. 03

    You get a written plan

    Within 48 hours: the recommended architecture, what it costs, how long it takes, and what I would cut if the budget is tight.

  4. 04

    You decide, no pressure

    Take the plan and hire whoever you want. If you would rather I built it, we scope it then — not before.

Questions

The things everyone asks, answered straight

Yes. I regularly work with founders who cannot read code. I explain trade-offs in business terms — cost, risk, timeline — and tell you plainly when a decision is a technical one you can safely delegate.

Usually LangChain or LangGraph with a vector store, FastAPI, and whatever hosting fits your scale — not whatever is trending. I will tell you when the boring option beats the interesting one.

Yes, and I prefer it. The Architecture Deep-Dive is designed so your own team can execute the plan. I only take on Build Partnerships where you genuinely want me to write the code.

Then I will tell you on the first call, and I will not take your money to prove it. Plenty of projects are better solved with a spreadsheet, a cron job, or a rule-based system.

The Clarity Call is free, so there is no risk there. For the Deep-Dive, if the written plan is not useful to you, say so within 7 days and I will refund it in full.

You pay when you book, through Cal.com. It takes a card or UPI, and the receipt is automatic. There is no invoice to chase and no payment after the fact.

I am based in India (IST) and regularly take calls with clients in the US and UK. Calls are scheduled in your timezone, not mine.

Still unsure? and ask me directly.

Experience

Professional Experience

A timeline of my roles and key contributions in AI and software engineering.

Jan 2026 – Present

Astraveda

Freelance AI Engineer

Jan 2026 – PresentRemote

Architected and deployed production-ready website and web application, maintaining full ownership of development, infrastructure, and hosting operations for enterprise petroleum analytics platform.

Key responsibilities
  • Engineered full-stack solution leveraging Next.js, Tailwind CSS, Shadcn UI, FastAPI, PostgreSQL, LangChain, and LangGraph to deliver scalable AI-powered features across client-facing and internal systems
  • Accelerated development velocity using Claude Code, Google Gemini, Google Flow alongside AWS Amplify, AWS Lambda, Supabase, Groq, and Mistral AI for robust cloud infrastructure and LLM orchestration
  • Delivered two core AI capabilities: "Ask Astra" multi-agent conversational chatbot and "Click Astra" OCR solution for automated document processing
Website·App
Next.js
Tailwind CSS
Shadcn UI
FastAPI
PostgreSQL
LangChain
LangGraph
AWS Amplify
AWS Lambda
Supabase
Groq
Mistral AI

Nov 2025 – Dec 2025

Nuevosol Energy Pvt Ltd

AI Intern

Nov 2025 – Dec 2025Head office Madhapur, Hyderabad, India

Engineered enterprise Agentic RAG chatbot enabling team members to query internal documentation through natural language, dramatically reducing information retrieval time and improving operational efficiency.

Key responsibilities
  • Architected multi-agent retrieval system using LangGraph orchestration, OpenAI LLMs, and ChromaDB vector database, implementing semantic search across company knowledge base
  • Deployed full-stack solution with Next.js frontend (Vercel) and FastAPI backend (Render), delivering production-ready chat interface with real-time response capabilities
LangGraph
OpenAI
ChromaDB
Next.js
FastAPI
Vercel
Render

Feb 2025 – June 2025

Prodigal AI

Agentic AI Intern

Feb 2025 – June 2025Remote, India

Worked on Dhanur AI, a cutting-edge video editing automation platform by Prodigal AI that transforms raw user video into polished YouTube/Shorts-ready content using Langchain framework.

Key responsibilities
  • Implemented Retrieval-Augmented Generation (RAG) pipelines integrated with vector databases for intelligent context retrieval of b-roll, filters and transition segmentation using ChromaDB
  • Reduced manual editing time significantly through intelligent automation, enhancing creator productivity
Achievements
  • ★ Awarded Intern of the Month in April 2025 for exceptional performance and innovation
Python
LangChain
ChromaDB
RAG
Vector Databases
FastAPI

Contact

Get In Touch

Interested in collaborating on AI projects or discussing opportunities? I’d love to hear from you.

Email
sar.abhijit2003@gmail.com
Phone
+91 9985727779
Location

India

Connect
githublinkedinwebsiteemail