Live five-week cohort · Starts Wednesday, October 14
Place your first AI trade on day one. Finish with an AI agent that researches and trades for you every day.
Use AI to research, scan, trade, and run an AI trading agent with real money.
Illustrative demo of the live cohort workflow.
See how a plain-language idea moves through review, approval, and monitoring.
Tools, brokers, and markets you'll work with
5 Live sessions with RobOne session a week, taught live and recorded.
3 hours A weekThe live session plus time to practise.
30 days Money-back guaranteeRequest a refund within 30 days if it is not for you.
We searched Reddit and X for the trading problems people complain about most. AI makes each of these much easier, and the cohort shows you how.
Tickers, news, levels, endless tabs. Traders on Reddit and X keep saying AI cut that grind from hours to minutes.
How we fix this
We teach you a Claude, Codex, Cursor, or BotSpot scan workflow that builds a short, broker-checked list so research stops eating your day.
The model sounds sure, then the number is stale or off by a few ticks. Close enough to trust. Wrong enough to hurt.
How we fix this
We show you how to connect to your real broker so prices, positions, and account value come from the account you trade, not a chatbot guess.
A short win streak is not proof. Traders keep asking the same thing: was that edge, or a fluke?
How we fix this
We show you how to backtest and run evals on an AI trading agent so you can tell a lucky streak from a process that holds up.
Broker connections drop, sessions log you out, and a lot of people still chart in one place and click the order somewhere else.
How we fix this
We show you how to use Claude, Codex, Cursor, or BotSpot so chart work and a broker-ready trade live in the same agent workflow.
Anyone can paste a prompt and get an indicator. The hard part is knowing what it measures, when it fails, and whether you would trade it with real money.
How we fix this
We teach you to generate, stress-test, and use indicators with Claude, Codex, Cursor, or BotSpot so you ship something you understand.
Platform script errors, outdated tutorials, and “too complex” walls show up the second you try to scan the market your way.
How we fix this
You build scanners in plain English with Claude, Codex, Cursor, or BotSpot, then verify names against real market data instead of wrestling platform code.
Multi-leg options are tedious to build, fill, and manage in broker UIs. Traders talk about partial fills and babysitting every leg.
How we fix this
You describe the structure in plain English, review the legs with AI, then send a checked order through your broker instead of clicking each leg by hand.
Research, scan, review a trade, repeat. In plain English. Not another maze of menus, scripts, and disconnected screens.
How we fix this
That is the core of the cohort: Claude, Codex, Cursor, BotSpot, and your broker working as one command seat you control.
Problem, then practice: Claude, Codex, Cursor, BotSpot, and your broker working together, with real money under your control.
This is a live five-week bootcamp taught by Rob Grzesik, CEO of Lumiwealth BotSpot. He helped build Voyager into a billion-dollar crypto company, and he built technology at Greystone, a mortgage lender, that handled over $3 billion in transactions. He also has a Master of Finance, 25+ years of coding experience, and 15+ years in financial technology.
You are not watching lecture videos. Each week you show up on Zoom for a hands-on class, build the workflow live, place a reviewed trade, and leave with a recording you can replay forever. You also join the community: work with other students, retake future classes, and keep new material we add going forward.
Week one gets your AI trading workspace working: Claude, Codex, Cursor, BotSpot, market data, and your broker. You place your first stock or ETF trade the same night.
Week two turns AI into a research desk that saves hours. You pull filings, financials, earnings, insider data, and news, then build scanners that surface the next candidates. Stop spending hours building watchlists. Stop guessing when you skip the work.
High-ticket cohorts work when the right people enroll. If this is you, keep going. If it is not, save your seat for someone else.
Want to check the fit? Ask the instructor directly.
Claude, Codex, Cursor, and BotSpot all talking to your broker on night one.
Filings, financials, earnings, insider data, news, and a scanner that surfaces the next candidates.
Use Claude with TradingView, then place reviewed trades beyond stocks and ETFs.
Research, decide, and place trades on a schedule inside guardrails you control. Under your command.
Turn a plain-English setup into rules that run without babysitting every click, still with you approving the system.
Watch a real Challenge lesson, then see how the five-week Bootcamp turns that workflow into your own AI trading agent.
Taught live by the founder. Not a hired instructor reading slides.
Rob Grzesik · CEO of BotSpot · Creator of Lumibot · Master of Finance, Schulich
In 2008 I blew up my own trading account. I had conviction, I had opinions, and I had no way of knowing whether any of it was actually true before I put money behind it. So I stopped guessing. I taught myself to backtest everything, then earned a Master of Finance at the Schulich School of Business at York University. When my father passed away I took over my family's accounts, and that same discipline grew them many times over. That money is still working today, and it is why I take this seriously.
After grad school I built a software company and sold it, which paid for a couple of years of travelling. Then I landed in New York at Greystone as VP of Fintech, where I built loan software that processed over $3 billion in transactions. After that I was one of the first engineers at Voyager Digital, building the trading and brokerage infrastructure from scratch and growing the engineering team at a company that reached a $1B+ peak market cap. I later served as VP of Software Engineering at Perch and have hired and managed 100+ engineers.
Along the way I created Lumibot, now nearly 2,000 stars on GitHub, personally built more than 1,000 trading systems, and have taught algorithmic trading and AI trading to 14,000+ students over the last seven years. What changed recently is that AI agents got good enough to do work that used to need a desk of analysts: read the filings, pull live data, run the analysis, and prepare the trade. Almost nobody knows what these tools can actually do yet. That gap is the reason this bootcamp exists.
Build live each week, ask questions on camera, and bring your own setup for review.
Join once and keep coming back. Retake later cohorts as many times as you want, and keep access to new material we add.
Every live session lands in your portal. Rewatch this cohort whenever you want, plus past classes and future sessions we add to the library.
The exact checklists, prompts, and working files used in class, so you can rerun the same workflow after the call ends.
Private working time on your own setup, your own account, and the trades you actually want to make.
A full month of the Pro plan so you can research, backtest, and run bots on the platform while the cohort is live.
Keep building with us after class. These run every week and come with your Pro month.
The strategies we have written and use ourselves, in code you can read, clone, and change.
Optional five-week competition. Choose paper or real money, use one fixed starting balance, and trade with AI.
30-day money-back guarantee.
Connect Claude, Codex, BotSpot, and your broker, place your first reviewed stock or ETF trade with real money if you want, and join the optional five-week trading competition.
Finish night one with a working AI trading desk and a real trade on the books. Real money is encouraged. You place the trade yourself.
Research another stock or ETF, write a short research note, and place a second reviewed trade for the competition before week two.
Use AI to research stocks and ETFs with filings, financials, earnings, news, and insider data, then build a market scanner that surfaces the next candidates.
Use AI to research stocks and ETFs in class, verify the evidence, and build a scanner so you are not hunting charts by hand.
Write a full research memo on a company or industry of your choice, finish or refine your own market scanner, and place another competition trade before week three.
Use Claude with TradingView, turn chart ideas into trades, and place a reviewed trade in options, crypto futures, or Polymarket.
Week two covered stocks, ETFs, and scanners. This week opens options, crypto, crypto futures, and prediction markets with Claude on TradingView.
Place one competition trade in options, crypto or crypto futures, or Polymarket before week four. Bring the chart, the reason, and the risk limits you used.
Turn one plain-English setup into a Python strategy, backtest it, improve it, and run the same rules automatically on your broker with real money under your control.
Leave with a simple Python strategy you understand: clear rules, a backtest you can read (a historical simulation of how the rules would have traded), and code that can place those same trades on your broker without you rewriting them every day.
Finish your own simple Python trading strategy that can trade automatically. Run your own backtest and bring the rules, results, and one real-money signal to week five.
Create an AI-powered agent that can research, decide, and trade automatically for you every day on a schedule with real money under your control.
Week four was a Python strategy with fixed rules in code. Week five is an AI agent that can think, research, and act in plain English inside limits you set, around the clock on your brokerage account with real money.
No formal homework. The bootcamp and competition wrap this week. Optional: keep your AI agent running with real money under your rules and note one improvement you want next.
Join the optional five-week leaderboard with paper or real money. Everyone uses one fixed starting balance and trades with AI.
Syllabus above is what you build each week. This is when the live cohort runs.
October 2026 AI Trading Course
Build it live, then keep every recording.
Try before you buy
Before you pay for the five-week bootcamp, join the free 7-Day AI Trading Challenge. Start one useful AI trading project at the live show, then keep improving that same project with light tips over the week. There are no daily assignments or required submissions. See how Rob teaches, then enroll in the full cohort if it fits.
Hear from students who took the course.
Joel
Santayan
Agim
LaToya
Anthony
Daniel
October 2026 AI Trading Course: the price goes up on set dates. Enroll earlier and pay less.
$1,997was $2,999
Until Sunday, October 4
$2,999
After Early Bird ends on Sunday, October 4 at 11:59 PM EDT
Early Bird: $2,999 $1,997 pay in full
Early Bird ends in
No coding experience required
Try before you buy
Before you pay for the five-week bootcamp, join the free 7-Day AI Trading Challenge. Start one useful AI trading project at the live show, then keep improving that same project with light tips over the week. There are no daily assignments or required submissions. See how Rob teaches, then enroll in the full cohort if it fits.
No. You direct Claude and Codex in plain English, and we build the workspace together. Deeper coding paths are optional if you want them later.
It means giving AI a trading job it can repeat: research the market, scan for a setup, prepare or place a reviewed order, and run again on a schedule inside rules you control.
You can use paper or real money. Early trades are reviewed before they go out. Later automation runs inside approval rules, size limits, alerts, logs, and a stop control that you set.
You start by using AI to research and place reviewed trades. By the end you have an AI agent that can research, decide, and trade every day inside guardrails you set.
Weeks one and two focus on stocks, ETFs, research, and scanners. Week three covers options, crypto and crypto futures with Bitunix, and prediction markets with Polymarket, using Claude with TradingView.
Five Wednesday evenings at 7:00 PM Eastern. Each live session is about 1 to 2 hours. Plan on about 3 hours a week total once you include practice outside class. Exact dates are listed in the Schedule section. Every session is recorded.
Rob demonstrates the workflow, you build alongside him, and the group reviews real setups and questions.
Every session is recorded and added to your portal. You keep the library, including past classes and future sessions we add.
Yes. Join once and you can retake later live cohorts as many times as you want. New material stays in your course library.
Join the optional five-week leaderboard with paper or real money. Everyone uses one fixed starting balance and trades with AI. Trade your way: run a strategy you build in class, place your own trades with AI research, or let your AI agent trade for you. Each week of class hands you a new AI skill to try in the competition, so you practice on real market decisions instead of homework. One class leaderboard tracks participants. Best return at the end of week five wins.
Yes. Join the free 7-Day AI Trading Challenge first. Start one useful AI trading workflow at the live show, then keep improving that same project with light tips over the week. Dates come from the live challenge schedule.
Pricing windows, dates, and amounts all come from the live promotion schedule for the current cohort. See the How pricing works section on this page for the windows that are active right now.
30-day money-back guarantee.
The next cohort starts Wednesday, October 14.
No coding experience required
Prefer to learn on your own schedule? Get the self-paced version for $499
Past performance does not guarantee future results. Automated trading involves risk of loss.
