Engagements
A record of talks, workshops, and panel contributions where machine learning methods for cryptocurrency market analysis were presented to technical and practitioner audiences.
Invite to speak
What Machine Learning Actually Does When Analysing Cryptocurrency Data
A plain look at the mechanics, for readers who want to understand the process rather than the hype
A plain-language look at how automated systems process cryptocurrency market data, and what has shifted in how these tools work over the past two years.
Read more
Machine Learning and Cryptocurrency Markets: A Grounded Overview for Cautious Readers
The technology is real, the limitations are equally real, and both deserve a clear explanation
If you have heard that computers now analyse digital currencies automatically, here is what that means in practice and why the technology is more limited than headlines suggest.
Read more
Three Specific Things That Changed in ML-Based Cryptocurrency Analysis Since 2023
Concrete developments worth knowing about, explained without unnecessary technical jargon
A focused look at concrete technical developments in machine learning applied to crypto markets, written for readers who prefer specifics over generalities.
Read more
The Data Work Behind Cryptocurrency Machine Learning Models
The preparation work that precedes every model output, and why it shapes everything that follows
Before any model produces an output, a large amount of unglamorous preparation work happens. Here is what that process looks like and why it matters more than most coverage acknowledges.
Read moreBefore and after
Participants arrive with general programming knowledge and leave with a working understanding of how sequence models process time-series data from crypto markets. The gap between those two states is measurable.
Each workshop session covers one concrete method - from feature engineering on OHLCV data to evaluating a trained LSTM against a naive baseline. Participants run the code themselves, not just watch it run.
Numbers reflect a single full-day workshop. Participants build a baseline model, a feature-engineered version, and a sequence model - each evaluated against real held-out data.
What a session covers
Talks are structured around a single applied problem - typically predicting a directional signal from crypto price data using a specific ML method. The audience follows along with code, not slides.
Sessions work best for technical audiences who already write code and want to see how ML methods behave on real, messy financial data rather than cleaned benchmark datasets.
Discuss a sessionEvery method is demonstrated by writing and running code during the session, not by showing pre-built outputs.
Sessions run as 90-minute focused talks or full-day workshops depending on the depth the event requires.
Both formats are supported. Remote sessions use shared notebooks so participants run the same environment without setup friction.