Hi, I'm Tim

I'm a cybersecurity professional who thought building a trading bot would take "a couple of weeks, tops."

That was nine months ago.

The "What If?" That Started Everything

Between IT contracts a few years back, I spent three months day trading. The first six weeks? Losses. Every. Single. Day. Brutal!

But I persevered (thanks to paper trading!) and started to get some green days. My "eye" was developing—I could finally see what made a good trade.

But life called. A new contract came through, my family was growing, and I took the safe route. I went back to cybersecurity.

That "what if?" question never left me.

Why I Built TradeBOT

Fast forward to March 2025. I wanted to gain experience with machine learning, I'm a hands-on learner—I need to build something real, not just read tutorials.

Combining trading (which I love) with IT infrastructure and machine learning seemed perfect. Even if the bot failed, I'd still achieve my goal.

I fired up Visual Studio for the first time in 10+ years. Paired with GitHub Copilot and Claude, I had my first XGBoost model trained and running in a couple of days.

I thought, "This is easy! Two weeks and I'll have a working trading bot."

I was spectacularly wrong.

Nine Months Later...

I've tested 50+ technical indicators. Trained countless XGBoost models. Built an entire platform called SPIRIT from scratch.

The results? Not what I expected.

Model performance was underwhelming. Different indicators barely moved the needle. YouTube made this look easy—it's not.

But here's the thing: I love it!
Yes it's an emotional roller-coaster but for some reason I keep coming back.

What I've Built (So Far)

SPIRIT Platform

This turned out to be the real achievement. SPIRIT is a modular Python trading platform that:

  • Keeps positions private (no more getting hunted by high-frequency traders triggering your stop-losses)
  • Can run in test mode with CSV data feeds for safe algorithm testing
  • Swaps APIs easily (different exchanges, different algorithms)
  • Includes a debugging harness to feed specific data and test algorithm responses in real-time

It's become a serious piece of software—way beyond my initial "quick project" vision.

My Background (Why I Can Build This)

I've spent 10 years in cybersecurity and even longer in enterprise IT. Early in my career, I earned my MCSA in SQL and Windows Server.

I've worked in large Security Operations Centers (SOCs), designed networks and firewalls, and managed projects using PRINCE2 and Agile methodologies.

All those years of gathering business requirements, presenting roadmaps, and running projects in Jira? Turns out they're incredibly valuable for managing a complex R&D project like TradeBOT.

I've been coding in Python for a year now (thanks to this project), and AI tools like Copilot and Claude have been absolute game-changers.

I have to thank my brother-in-law who just completed his degree in computer science this year —collaborating with him early on in the project was invaluable.

The Honest Truth

I'm not sure where I am with this project anymore.

What started as a "1-2 month learning project" has become a full-scale software application with serious R&D.

After nine months of daily work, I feel like I'm close... but it's still not working profitably.

Why This Blog Exists

After nine months of work, I wanted to document where I am, make it easier to collaborate and provide a more honest view on tradebots.

YouTube videos make trading bots look easy. They're not. I've gone down plenty of wrong paths, and maybe sharing my journey can help others avoid the same mistakes.

I'm also considering opening SPIRIT up as a platform for others to build algorithms on. Maybe releasing the work I've done will help other traders and developers.

And honestly? This blog might generate support for the project—donations, subscriptions, or just interest from people who want to follow along.

I'm not charging for content yet. Right now, I just want to share the work and see if others are interested. If there's a following, maybe paid subscriptions or access to the PostgreSQL training data server could come later.

What You Can Expect

Brutal honesty. I'll share the wins and the losses.

Real code. Not just theory—actual algorithms, models, and infrastructure.

Technical depth, but accessible. I'm technical, but I don't always get it right. I'll explain things clearly and admit when I'm wrong.

The journey, not just the destination. The highs and lows have been incredible. If this project works, it'll be the best thing I've ever built. If it doesn't... well, I still learned at lot.

Join the Journey

I'm getting to the point where I want to open this project up to more people.

It's been my baby, so I feel protective. But more brainpower might mean more progress.

If you're interested in:

  • Algorithmic trading
  • Machine learning for finance
  • Building trading infrastructure
  • Following a real R&D project in real-time

...then stick around. Subscribe for updates. Drop a comment. Support the project if you find value.

Let's see if we can make this work.

Want to support TradeBOT?

Right now, the best way is to subscribe (it's free) and share the content if you find it valuable. Donations and paid tiers may come later if there's interest.

Want to collaborate?

Reach out via email: Tim@tradebot.live.com