I Wrote a Book on AI-Powered Nifty Trading — And Why It's Not Another "Get Rich Quick" Guide
Three months of research, 47 sections, 151 pages, and one NISM certification later, my first book is finally live on Amazon KDP.
Option Trading with AI: XGBoost, Transformers & Quantized Models for the Retail Nifty Trader is out.
If you're a retail trader in India who has ever stared at the Nifty option chain and felt overwhelmed by noise, hype, and contradictory Telegram alerts — this one's for you.
The Problem I Kept Running Into
I trade Nifty. I also code. And for the longest time, these two parts of my life lived in separate silos.
Every time I searched for "AI trading India," I found either:
- Academic papers that assumed HFT infrastructure I'll never have
- Instagram gurus promising "AI-strike-selection bots" for ₹999
- GitHub repos that copy-pasted US equity models onto Nifty data without noticing the structural differences (expiry cycles, tax slabs, circuit filters)
The market is not a generic dataset. Nifty options have a 52-week expiry calendar, max-pain dynamics, and an underlying that behaves differently from SPY or NASDAQ. A model trained on US equities breaks here.
So I stopped looking for a perfect open-source project and started building one — for my constraints, my phone, my capital, my time zone.
That experiment became the book.
What the Book Actually Covers
I don't believe in "chapter fluff." Every section is meant to be read, coded, and critiqued.
Part I — Future of Trading
We talk about what's actually changing in markets: sentiment pipelines, transformer attention on filings, and why "AI replacing traders" is the wrong framing. The leverage is in the process, not the prediction.
Part II — XGBoost & Walk-Forward Validation
Not just model.fit(). I walk through why ranking anomalies matters for Nifty, how to build time-series splits that respect expiry calendars, and how to avoid the overfitting trap that kills 90% of retail ML experiments.
Part III — Optuna on Android/Termux
Yes, you can run hyperparameter optimization on a ₹15,000 phone. I show the exact setup: SQLite storage, nightly trials, walk-forward metric, no cloud GPU needed.
Part IV — Quantization (FP32 → INT8)
This is the chapter I wish I had two years ago. ON-device models, private inference, lower latency — everything a retail trader needs when data charges and broker APIs are the bottleneck.
Part V — Research Ethics & Launch Checklist
A practitioner's handbook for the boring but critical stuff: disclosure, SEBI compliance, citations, draft review, and a launch checklist. Because real work is 80% process and 20% signal.
Why I'm Pointing Out the Restrictions
This book has a SEBI-style disclaimer on every chapter. It says, clearly:
This book is for educational and informational purposes only. It does not constitute SEBI-registered investment advice.
I am NISM Series-XII certified — that's a real credential from the Securities Markets Foundation, recognized by SEBI. But certification is a floor, not a ceiling. It means I know the regulatory boundaries. It does not mean I can predict Friday's expiry close.
Every trading idea in the book is educational. Every model is a research artifact, not a send-button bot.
Who Should Read This
- Retail Nifty traders who want a repeatable scan-and-rank workflow
- Data folks curious about where ML actually fits in an Indian options book
- Anyone who has been burned by "AI + trading" hype on social media and wants the boring, honest version
If you're looking for a holy grail indicator, this book will disappoint you.
If you're looking for a structured, research-first process — you'll find it here.
The Numbers
- Title: Option Trading with AI: XGBoost, Transformers & Quantized Models for the Retail Nifty Trader
- Pages: 151
- Trim: 6×9 in (paperback) + Kindle eBook
- Price: $9.00 USD (Kindle), paperback available
- ISBN: 979-81-88223-79-3
- Author: Shakti Tiwari — Nifty Option Trader, Research Analyst & XGBoost Expert; NISM XII Certified
Where to Find It
- Amazon Kindle: https://www.amazon.com/dp/B0XXXXXX
- Amazon India: Will auto-appear on Amazon.in after publishing
- GitHub: https://shaktitiwari715-ai.github.io/shakti-tiwari-nse — syndicated chapters + images
Final Thought
Writing a book doesn't make you a guru. It means you cared enough to organize what took you months of mistakes to learn.
Read it. Critique it. Run the code. Break the models. That's how we get better — together.
Happy trading, and may your walk-forward metrics always beat your training accuracy.
Shakti Tiwari
@shaktitiwari
Shakti Tiwari — Nifty Option Trader, Research Analyst & XGBoost Expert
