About Tovrask

Where machine learning meets investment practice

Tovrask is a Grand Falls-based online education platform built around one specific subject: how automated systems are reshaping portfolio management and financial decision-making.

View the learning program

A platform built around one focused problem

Most financial education covers broad territory. Tovrask was set up in 2018 to do the opposite — concentrate entirely on machine learning as it applies to investment automation. The curriculum addresses model selection, backtesting methodology, risk-adjusted signal generation, and the practical limits of algorithmic approaches in live markets.

Who the sessions are designed for

Participants typically arrive with some background in either finance or programming — rarely both. The webinars are structured to build shared vocabulary first, then move into technical implementation. Sessions run live with Q&A built in, not appended as an afterthought. Recordings are available but the live format is where most of the practical discussion happens.

6+ Years active Growing
40+ Live sessions delivered Ongoing
1 Subject, covered in depth Focused
Live webinar session on machine learning in investment automation
Live broadcast format — Grand Falls, NB

How sessions are structured

Each webinar follows a consistent format: framing the problem, walking through a working implementation, then opening the floor for specific questions. The goal is not inspiration — it is working knowledge that participants can apply.

Instructor presenting machine learning model architecture during live session
Format

Live instruction with technical depth

Sessions run between 90 minutes and two hours. The first segment covers theory and context — why a particular method exists, what problem it was designed to solve, and where it tends to break down. The second segment moves into code or data, depending on the topic. Participants are encouraged to follow along in real time.

Instructors are practitioners, not generalists. They work with the tools and datasets they discuss. When a model fails in a demonstration, that failure becomes part of the lesson rather than something to edit out of a recording. This approach reflects how machine learning actually behaves in investment contexts — inconsistently, with results that depend heavily on implementation choices.

Participant reviewing model output during interactive webinar exercise

Interactive Q&A

Questions are answered during the session, not collected for later. Instructors adjust the pace based on what the room needs.

Manual Automated ML The shift this platform teaches — from discretionary to model-driven investment logic

Materials & resources

Code samples, datasets, and reference documents are shared with registered participants after each session. No paywalls on follow-up materials.

Post-session access

Recordings remain available for a defined period. The platform does not replace live attendance with passive video — recordings supplement, not substitute.