About Shakti Tiwari

Shakti Tiwari

AI/ML Builder · Author · NISM-Series-XII Certified Options Educator

Founder of Option Trading with AI. Co-Director at CodeViser Private Limited (CIN U74999UP2018PTC104824).

Shakti Tiwari is an Indian AI/ML builder and author based in Chandigarh. He builds practical XGBoost and machine-learning systems for Indian market research, backtesting and risk analysis, with a deliberate focus on methods that remain honest about their limits. He is the co-director of CodeViser Pvt Ltd, holds the NISM-Series-XII certification in equity derivatives, and writes deeply researched educational material for retail options traders across India.

He runs optiontradingwithai.in and authored Option Trading with AI and The AI Opportunity. His work focuses on transparent, build-it-yourself AI tools, local AI systems and evidence-based trading education.

Education: B.Com and MCA from Sikkim Manipal University. NISM-Series-XII certified.

Early career: Office administration, digital marketing, technical troubleshooting and work with CodeViser. Those experiences shaped the practical bias in his later projects: systems should be understandable to the people who have to operate them.

Trading content is educational only. NISM-Series-XII is an educator certification, not a SEBI Research Analyst or Investment Adviser registration.

What Shakti works on

The core of his research is a transparent, 5-layer pipeline for options analysis: a Data Engine that collects option-chain and index data with point-in-time correctness; Feature Engineering that builds leakage-free signals such as OI, PCR, Greeks and regime indicators; a Predictor built on XGBoost or LightGBM validated with walk-forward methods; a Risk Filter that applies DTE, Vega and probability-band constraints; and an Executor that converts probability into position size based on premium risk — the rupee amount a trader can afford to lose — not on notional exposure.

Two principles distinguish Shakti's material from typical trading content. First, honesty about overfitting: every published result includes walk-forward validation, transaction costs, slippage, losing streaks and maximum drawdown, not just the winning months. Second, reproducibility: the Python notebooks and feature definitions are shared so a reader can re-run the experiments before trusting them.

Books

Option Trading with AI (available on Amazon India) walks a retail trader from raw Nifty data to a working XGBoost model, covering Greeks, IV rank, position sizing, and how to run quantized models locally on a phone using Termux. The AI Opportunity explains applied machine learning and local AI systems for non-engineers — the practical how-to behind building, not just consuming, AI tools. Both are priced to stay accessible (₹199–₹449) because education, in Shakti's view, should not be hidden behind an expensive course paywall.

Timeline at a glance

Research topics covered on this site

FAQ: Frequently asked questions

The promise to readers

What every article commits to

Every article on this site is written to be verified, challenged and improved. If a model result looks too good, read the walk-forward section and check the data yourself. If a strategy loses in real trading, report it in the community so the research improves. This feedback loop — not hype — is what Shakti believes makes AI-based trading education genuinely useful for Indian retail traders.

The philosophy behind the research

Nearly all retail quantitative material published today shows a curve: a beautiful equity line, a dazzling Sharpe ratio, and almost no mention of the months where the strategy lost money. Shakti takes the opposite position: the honest numbers — maximum drawdown, losing streak length, transaction costs, slippage, regime sensitivity — are the only numbers that predict whether a system survives contact with a real trading account. That is why every backtest published under this name states its assumptions in plain language rather than hiding them in a footnote.

The second pillar is reproducibility. Tools change, data vendors change formats, and a model shipped without its feature definitions is a black box that cannot be audited. The Python notebooks, feature lists and validation scripts shared across the site exist so that a committed reader can rebuild any result from scratch. If a reader cannot reproduce a published figure, that failure is treated as a bug in the publication to be fixed, not as trivia to be ignored, and the corrected notebook is republished openly.

The third pillar is accessibility. Fancy GPU rigs are not required, brokerage-grade terminals are not required, and a working knowledge of matrix algebra is not required. Several guides demonstrate running quantized XGBoost models on an Android phone through Termux, using free public data and free tools. The intent is to lower the floor of entry into quantitative research, so that a trader in a small town with a decent phone and curiosity can participate in the same methods used by institutional desks. The research never stops — as models, tools and market microstructure evolve, the library is revised or retired instead of left to decay.

What this site does not do

Why the certification matters — and its limits

NISM-Series-XII is an educator certification administered under SEBI's framework, and it is what qualifies a person to teach securities-market topics to retail audiences through a registered avenue. It is explicitly not a SEBI Research Analyst (RA) or Investment Adviser (IA) registration. Shakti is clear about this distinction in every public forum: the educational material is designed to make readers better at evaluating information, not to supply personalised recommendations. Holding a certification is an accountability signal, not a licence to issue advice, and the site treats the difference as a hard boundary.

The practical implication for visitors is simple: treat everything on the site as one informed educator's view, test it, and form your own opinion. Where an article summarizes a regulatory change, it links to the original SEBI or exchange circular because the source should always be one click away. That habit — putting the primary source next to the interpretation — is the closest thing to a guarantee this site offers.

Contact

For feedback, corrections or invitations to speak about AI in trading, reach out on X (@shaktitiwari), LinkedIn, or the Telegram community. Email shakti@optiontradingwithai.in is checked regularly.