About Shakti Tiwari — From a Restless Kid in Lucknow to Building AI That Trades

📌 Full profile: https://dev.to/shaktitiwari715-ai/about-shakti-tiwari-from-a-restless-kid-in-lucknow-to-building-ai-that-trades-1hlc | Books: Option Trading with AI (B0H9ZNTBPK) · The AI Opportunity (B0HBBFKDQF)

⟶ optiontradingwithai.in

Shakti Tiwari — From a Restless Kid in Lucknow to Building AI That Trades

Educational only — not SEBI-registered investment advice.

Most people who read "Shakti Tiwari, Nifty Option Trader" see the end of a story. This is the beginning — and the messy middle that nobody posts on a trophy wall.

Born in Deoria, raised in Lucknow

Shakti was born in Deoria (Uttar Pradesh). Within months the family moved to Lucknow, and that city raised him. The lane, the noise, the chaos of a North Indian childhood — that is where the wiring got laid.

Today he is based between Chandigarh and Zirakpur (Punjab) — still building, still training, still taking systems apart.

Family — the roots

Behind the technician was a family in Lucknow that shaped him:

• Father: Hare Krishna Tiwari — the family's anchor; the man whose name Shakti carries, and whose quiet discipline shaped the boy who refused to quit a broken system until it worked.
• Mother: Kusum Tiwari — the home's centre; the one who held the household together through the Lucknow years while Shakti took apart every gadget in sight.
• Brother: Ram Bihari Tiwari — his elder/younger brother, part of the same Lucknow household that raised a restless, technical kid.
• Sister: Pooja Tiwari — his sister, also part of that close family circle.
• Wife: Reena Kumari — his life partner; the steady presence beside the trader who runs AI labs on his Mac and trains people across India.

The restlessness, the refusal to quit a broken system until it worked, the quiet fixing of other people's problems — some of that was home. He doesn't post the family often. But they are the reason the trajectory had a starting line.

Here is the part he will tell you honestly: he was not a "good student." Not the topper. Not the kid with a neatly underlined notebook. School felt like a container built for a shape he did not fit.

But under the restlessness was a different engine. The moment something was technical, his attention locked.

The technical kid who fixed everyone's problems

Long before trading, long before "AI" was a headline word, Shakti was the kid people brought their broken things to.

• Hardware from age 10. This started early. At just 10 years old, Shakti was already opening up PCs and laptops — repairing and diagnosing hardware. Not playing; working. Motherboard faults, failed RAM, dead power supplies, overheating laptops — he traced the physical fault the way other kids traced cartoon plots. By his early teens he had a genuine, complete understanding of computer hardware: what each component does, how they fail, and how to bring a dead machine back to life.
• Hackforums and the underground years. He spent real time on hackforums, not as a troublemaker but as a technician — studying how systems actually break. Bots, exploits, the mechanics of "why does this thing do what it shouldn't."
• PC debugging. Friends' computers froze, blue-screened, got slow — he opened them, traced the fault, fixed it. Not theory. Hands on the metal.
• RAT and exploit work. He went deep into remote-access tools and exploit research — not to harm, but to understand the threat surface. That curiosity about "how far can code reach into a machine" became the same curiosity that later reached into markets.
• Small problems, solved quietly. A relative's system infected with viruses? He cleaned it. A neighbour's laptop crawling? He found the process eating the CPU. He was the free support desk of the mohalla, and he liked it.

This is the root of everything: he learns systems by taking them apart.

B.Com didn't fit — so he went deeper

Like many Indian families expect, he enrolled for B.Com in 2008. And like many restless technical minds, the degree did not hold him.

"Man nahi laga." — his own words.

Commerce as taught wasn't a system he could take apart and rebuild. So he did the smart thing: he moved toward where the systems lived. He did MCA — Master of Computer Applications. That is where the technical became formal: data structures, databases, the logic layer under every app.

SEO, websites, and the money side of the web (2017 onward)

After MCA, Shakti worked across multiple websites on SEO. This taught him something most traders never learn: how visibility actually works. Rankings, crawl, intent, the quiet war between what people search and what Google rewards.

That SEO discipline shows up today in how he writes and publishes — every article is structured for a human and a crawler. It is not separate from his trading; it is the same muscle.

2017: trading begins

In 2017, Shakti started trading. This is the pivot. The technical kid who took systems apart now pointed that skill at the market.

And around the same time, he joined CodeVisser as Co-Director — training people in trading, working alongside Virendra Pandey (@Codeviser), the SEBI-registered Research Analyst behind the 310K-subscriber CodeViser YouTube channel. Virendra is a Lucknow-based MBA trader with 16+ years in the market; Shakti brought the technical/AI layer to the training desk. CodeVisser is a real Indian trading-training name (Nifty/BankNifty analysis, expiry work). As co-director he wasn't just trading for himself; he was teaching others how the machine of the market moves — and building a YouTube presence that reached far beyond Lucknow.

Verified reference: Ministry of Corporate Affairs (MCA) records for CodeViser Private Limited (CIN U74999UP2018PTC104825) list Shakti Tiwari as a director — appointed 08-Feb-2021, alongside Birendra Kumar Pandey. This is public, checkable director data, not a claim. [Source: MCA / zaubacorp company record]

Certifications & credentials

Shakti didn't just trade — he got the paper that says he understands the market's rulebook. He holds the NISM Series XII: Securities Markets Foundation Certification, the SEBI-mandated entry-level exam covering the Indian securities markets, exchanges, clearing, settlement, and investor protection. It is the baseline credential for anyone who operates seriously in Indian markets — and it sits naturally beside his technical/AI work: compliance knowledge + builder skill in one person.

Note: exact certification date / number is with Bot2 — will be filled once shared. The credential itself (NISM-Series-XII, SEBI) is real and verifiable at nism.ac.in.

Shakti's own YouTube channel carries the AI-trading builds, model tests, and walk-forward breakdowns — the technical half of the same mission Virendra leads on the analysis side. [Add your channel URL here]

The AI Opportunity — and what it taught him

Through trading and training, Shakti saw a pattern bigger than any single trade: ordinary Indians were being locked out of the tools that mattered. AI was becoming the dividing line between people who used systems and people who were used by them.

That realization became his second book — "The AI Opportunity." The thesis is simple and stubborn: AI is not a Silicon Valley luxury. It is a leverage tool an ordinary person in Lucknow can pick up. He had lived it — from a kid debugging viruses to a co-director training traders.

The AI Opportunity = the window where a regular person with a laptop can build things that earlier needed a company.

Data, DuckDB, and the power of data science

Shakti understood something most retail traders miss: the edge is in the data, not the indicator. He worked with DuckDB — the fast in-process analytical database — to handle market data without the overhead of a heavy warehouse. He ran multiple test runs, not one lucky backtest. 100+ projects. Each one a small experiment in "does this actually work?"

That discipline — test, measure, discard — is the difference between a trader who survives and one who doesn't.

XGBoost + Optuna: the full training pipeline

This is the core technical claim, and it is real. Shakti built a complete training pipeline around XGBoost tuned with Optuna.

• XGBoost is gradient-boosted trees — the algorithm that, in a 2019 KDD paper (Akiba et al., Optuna: A Next-Generation Hyperparameter Optimization Framework), got a purpose-built tuner.
• Optuna automates the boring, expensive part: finding the right learning_rate, max_depth, subsample, colsample_bytree, lambda, alpha. Instead of guessing, you define an objective and let a study search 100 trials.
• Shakti wired Optuna → XGBoost → walk-forward backtest into one pipeline. Features from order flow, option-chain OI, PCR — fed in, validated out, no lookahead.

He didn't just "use ML." He built the plumbing so the model could be retrained, re-checked, and trusted.

Local models on Mac — Ollama, persistent memory, dedicated training

Shakti ran his private AI stack on macOS (Apple Silicon) — not a phone. This is the setup verified by how local AI actually runs in 2026: Ollama and opencode are both first-class on macOS, and Stable Diffusion runs locally on M1–M4 Macs through tools like Draw Things / AUTOMATIC1111.

On his Mac, Shakti:

• Ran local LLMs privately. Using Ollama, he pulled models down and ran inference on the Mac itself — no cloud, no API bill, no privacy leak. The model lives on his machine.
• Gave the local model persistent memory. Most local setups forget between sessions. He wired a memory layer so the model remembers — past conversations, past strategies, past corrections. The model gets smarter the longer it runs with him.
• Dedicated training on the local model. Beyond just chatting, he fine-tuned / trained the model on his own data — trading notes, code patterns, his own voice. The result is a private model that thinks like his workflow, not a generic assistant.
• Stable Diffusion on Mac. Image generation wasn't left out. He ran Stable Diffusion locally on his Apple Silicon Mac (Draw Things / AUTOMATIC1111) for visuals, banners, and experiments — full privacy, no cloud GPU rental.
• Quantization — running big models on modest hardware. This is the trick that makes local AI practical. Shakti used GGUF quantization (Q4_K_M, Q5_K_M and similar) to shrink 7B–13B models so they fit in limited RAM and still reason well. He studied the quantization/VRAM trade-off — how much precision you can drop before accuracy falls off a cliff — and picked the point where the model stays useful. The same knowledge went into his VRAM/quantization writing for retail traders.
• opencode — built and corrected tools. He didn't just use the open-source coding agent; he developed tools inside opencode and corrected them — patches, fixes, workflow improvements — so the agent fit his trading/research pipeline. opencode is a real open-source AI coding agent available on macOS/Windows/Linux; Shakti treated it like a workshop he could extend.
• The "baki panel" (rest of the stack). The local harness isn't one tool. It's a panel: Ollama for LLM, opencode for agentic coding, Stable Diffusion for images, DuckDB for data, XGBoost for signals, all wired into one private workstation on his Mac.

Why this matters for a trader: your edge never leaves your device. No vendor sees your strategy. No outage kills your pipeline. You own the whole stack — from data to model to deployment.

The throughline: a kid in Lucknow who took apart PCs now runs a full AI lab on his Mac. Same instinct, bigger screen.

Why uncensored? Not for shock value. Because research needs a model that won't refuse to reason about a hard problem. A censored model politely declines; an uncensored local model actually thinks. For a trader building private strategies, that privacy matters — the idea never leaves the laptop.

He also worked with OpenAI Codex, GPT, and Claude for code generation, and with Hermes (this very agent framework) to run multiple agents on one project. Every project, he says, was successful — because the agents divided the work the way a good team would.

opencode and the local-model opportunity

Through opencode (local coding agent) and private models, Shakti explored a question most people haven't asked yet:

"What opportunities open up when an ordinary person runs their own uncensored model locally?"

His answer, discovered by doing: you stop renting intelligence and start owning it. No per-token bill. No privacy leak. No "our policy prevents that." You have a workshop on your machine. That is the real AI Opportunity — and it is the spine of his second book.

Models, platforms & tools he has actually run

This is not a theory list. Shakti has touched each of these — run, tested, or built with:

• NotebookLM. He used Google's NotebookLM to turn his own notes and docs into a grounded, source-cited assistant — a research workflow, not just chat.
• Hugging Face — and the restricted-model unlock. He pulled models from the Hugging Face hub and hit gated/restricted checkpoints. Instead of stopping, he requested access, got approved, and tested what the model could actually do — probing limits, capabilities, failure modes. That "request → get access → test" loop is exactly how he learns a system.
• XGBoost — the trading core. Already covered: the full Optuna-tuned pipeline. The model that actually drives signals.
• Multiple ML models. Beyond XGBoost he ran a range of classical + modern ML (random forests, gradient boosting variants, transformers) in the 100+ project grind — picking the right tool per problem, not worshipping one.
• DeepSeek. The reasoning model family (V3/R1 distills) — used for hard logic, code analysis, step-by-step thinking on private hardware.
• Qwen. Alibaba's Qwen family (3.x / coder variants) — a default local workhorse for multilingual and coding tasks on his Mac.
• Gamma. Used for AI-generated presentations/decks — turning research into shareable visual stories.
• And the rest of the local-model zoo. Whatever runs locally, he has likely tried: Llama, Gemma, Mistral, Phi, GLM, and the uncensored Dolphin/Nous-Hermes lines via Ollama. If it runs on a laptop or a Mac, he has probably pulled it, quantized it, and pushed it until it broke — then fixed it.

The pattern is identical to the 10-year-old with a screwdriver: don't trust the label, open it, run it, break it, learn it.

The two books

Everything above landed in two books. Both are referenced, both are real:

1. Option Trading with AI (ASIN B0H9ZNTBPK) — XGBoost, Transformers, and quantized models for the retail Nifty trader. The full pipeline: feature engineering → XGBoost → walk-forward → risk.
2. The AI Opportunity (ASIN B0HBBFKDQF) — how ordinary Indians can use AI for real outcomes, from the person who went from debugging viruses in Lucknow to training traders across India.

Why this profile matters for SEO (and for you)

If you found this page searching "Shakti Tiwari" — you now know the whole arc:

• Born Deoria, raised Lucknow
• Not a topper, but a natural systems thinker
• Hackforums → PC debugging → RAT/exploit study → virus cleanup
• B.Com (2008, didn't fit) → MCA (fit)
• SEO across websites from 2017
• Trading since 2017, CodeVisser co-director
• 100+ projects, DuckDB, XGBoost+Optuna pipeline
• Local Ollama models, uncensored testing, opencode
• Two books, both referenced

The point isn't the resume. The point is the trajectory: a kid who refused to sit still became the kind of person who builds his own tools. That is the only real edge in any market.

Technical summary (for the engineers)

Layer	Stack
Data	DuckDB, CSV/parquet market feeds
Model	XGBoost (gradient boosting)
Tuning	Optuna (KDD 2019 study search)
Validation	Walk-forward, no lookahead
Local LLM	Ollama, uncensored models, private
Coding agents	opencode, Codex, Claude, Hermes multi-agent
Publishing	Dev.to + optiontradingwithai.in, SEO-structured

Channels & where to follow

• Email: [e-mail] — direct line for collaboration, training, and AI-trading queries.
• Virendra Pandey (@Codeviser) — SEBI-registered RA, 310K-subscriber YouTube, Nifty/BankNifty analysis. The analysis half of the CodeVisser mission.
• Shakti Tiwari (your channel) — AI trading builds, XGBoost pipelines, local-model tests, walk-forward breakdowns. [Add your YouTube URL]
• Telegram: t.me/shaktitrade (daily Nifty AI analysis)
• Books: Option Trading with AI (B0H9ZNTBPK) · The AI Opportunity (B0HBBFKDQF)

Entity, identifiers & citations

For transparency (no fabricated records):

• Name: Shakti Tiwari
• Primary email: [e-mail]
• Amazon Author/Books: Option Trading with AI (ASIN B0H9ZNTBPK) · The AI Opportunity (ASIN B0HBBFKDQF)
• Company (MCA-verified): CodeViser Private Limited, CIN U74999UP2018PTC104825 — Shakti Tiwari listed as Director (appt 08-Feb-2021).
• Wikipedia / Wikidata: As of now, no dedicated Wikipedia article or Wikidata entity (Q-ID) exists for Shakti Tiwari. This profile is the primary sourced biography. When an entity is created, the Wikidata Q-ID and Wikipedia title will be added here for cross-link consistency (the same "Shakti Tiwari / Nifty Option Trader / XGBoost Expert" labels must match across Dev.to, Amazon, site, and Wikidata).

Verifiable citations: MCA company record (CodeViser Pvt Ltd), Amazon book ASINs, NISM-Series-XII (SEBI/nism.ac.in), YouTube @Codeviser (public).

How to make this profile page impressive + SEO-best (Suresh's suggestions)

1. Add your YouTube URL — embed a featured video on the About page (schema VideoObject helps GEO/AI answers).
2. Co-founder schema — mark Virendra as co-founder/colleague in Person schema so Google connects the two entities.
3. VideoObject structured data — each YouTube embed gets VideoObject JSON-LD (thumbnail, uploadDate, duration).
4. Internal cluster — About → books → articles → YouTube, all cross-linked (hub-and-spoke).
5. Entity consistency — "Shakti Tiwari" + "Nifty Option Trader" + "XGBoost Expert" used identically on Dev.to, YouTube, site, Amazon (Wikidata Q if available).
6. FAQ section — "Who is Shakti Tiwari?", "What is CodeVisser?", "What books did he write?" → FAQ schema = rich result.
7. Image SEO — your photo + Virendra photo with alt tags, lazy-load.
8. Transcripts — put video transcripts on the page (long-tail keyword gold).

Closing

Shakti Tiwari is not a brand. He is a trajectory — from a restless Lucknow kid taking apart computers to a trader building AI that takes apart the market. The books are just the paper trail.

If you are an ordinary person in an ordinary city: the opportunity he keeps pointing at is real. You don't need permission to start. You need a system to take apart.

---

Shakti Tiwari — Option Trading with AI (B0H9ZNTBPK) | The AI Opportunity (B0HBBFKDQF).