The engineering they won't teach you on YouTube. A 4-week build-along where you ship a production Alpaca bot — same architecture I run on my own paper account, every bug fix included.
A 72% win-rate backtest means nothing if your bot dies on a network blip at 10:30 AM. Most courses stop at "look at this equity curve." That's where the real work starts.
The Mac sleeps mid-session. Your bot lags 2 hours behind, takes zero trades, the day is dead. You don't know it happened until you check the dashboard.
The cancel-order endpoint returns HTTP 204 (no body). Your `r.json()` crashes. The whole bot dies mid-trade with stops half-replaced.
One urllib3 ConnectionError during a signal check and the whole process exits. You discover at 4pm that you never traded.
OPG orders silently expire. Stop legs go to "held" status, not "new." Your dashboard says no stop — except there is one, you're just querying wrong.
Every bug in this course is a bug I hit in production. Every fix is the fix I shipped. I'll show you my live dashboard, my green days, my crash days, and the exact code that handles both.
Backtest of the School Run v8b strategy (included in the course). Past performance ≠ future results. Real-world execution typically 60–75% of backtest after slippage and missed fills.
school_run_live.py — the same hardened script I run. With every fix from 7 production bugs, fully commented.
Flask + Chart.js single-file dashboard. Live positions, ATR trail state, closed trades, equity curve, system health.
6 years of cleaned 15-min data for 19 symbols. Variant tester so you can swap entries, exits, filters and re-run in seconds.
Every production bug I've hit, the symptom, root cause, and the patch. PDF reference you'll actually use.
4 weeks, 8 lessons per week. Screen-recorded, no fluff. Watch in 15 hours. Build your bot in parallel.
Discord for the cohort. I'm in there daily for 4 weeks. After that, lifetime access — alumni help each other.
Most strategies look amazing in backtest and die in real life. We'll dissect why — and build School Run v8b, an ORB strategy that survives 6 years of slippage, gaps, and regime shifts.
This is the week most courses skip. Alpaca's API has 47 ways to ruin your day. We'll walk every one.
Account.unrealized_pl returns None. The activity log lies. Yesterday's equity isn't yesterday's equity. We'll build a dashboard that surfaces the truth, not the convenience fields.
Your laptop is not a server. The final week takes you to the cloud — Hetzner, systemd, Tailscale, the works. You'll deploy, recover from a simulated crash, and walk away.
No subscription. No upsell. You buy it once and own everything — code, data, dashboard, future updates.
You need to be able to run a Python script and use a terminal. If you've ever installed a Python package with pip, you're qualified. If you've never written code, this isn't for you — try a "no-code" tool first.
Maybe. Maybe not. The strategy has a 72% backtest win rate over 6 years; my live paper performance has been mixed (we just finished a +$248 week and a flat week). Real-world performance is typically 60–75% of backtest. Anyone promising guaranteed returns is lying. I'm teaching you to build the engine — what you do with it is up to you.
For real money, yes — under $25K you hit PDT limits that skip most signals and cripple returns. For learning, no — Alpaca paper accounts are free and unlimited. Build it on paper first. Move to real money only when you've seen 60+ days of live behavior.
$97 attracts people who won't finish. $1997 prices out the people I actually want to teach — devs and quants who like building things. $497 sits where the real work meets the real budget.
No. I'm not an investment adviser. I'm a builder teaching other builders how to build. The course teaches engineering, not financial advice. The strategies are examples — you choose what to run, how to size it, when to stop.
30-day money-back guarantee. Email me, refund processed. I'd rather you not be the customer who hates me. No fine print.