What We Teach
Structured, hands-on programmes that take you from raw data to working models - built around real cryptocurrency datasets, not toy examples.
Programmes offered
Each programme at jynovue is structured around a specific skill gap - not a broad topic. Participants work with Python, pandas, scikit-learn and time-series libraries from the first session, applying each concept to historical cryptocurrency market data before moving to the next.
The format is remote and self-paced within a fixed cohort schedule. You get access to recorded sessions, a shared notebook repository, and a private group channel where participants post their work and get feedback from instructors. Progress is measured by what you build, not by quiz scores.
Programmes are designed for people who already have basic Python familiarity and want to apply machine learning methods to a specific domain. We do not cover general data science theory - the focus stays on cryptocurrency price and volume data throughout.
Data Preparation for Crypto Signals
A focused programme on cleaning and structuring raw exchange data. Participants work through missing value handling, resampling OHLCV data, and building feature sets that machine learning models can actually use.
- Handling gaps in tick and candle data
- Normalisation strategies for price series
- Feature engineering from volume and order flow
- Exporting clean datasets for modelling
Supervised Models on Market Data
Participants train classification and regression models on historical cryptocurrency data. The programme covers model selection, cross-validation on time-series splits, and interpreting outputs without overfitting to recent history.
- Decision trees and gradient boosting on price data
- Walk-forward validation setup
- Feature importance and model debugging
- Avoiding data leakage in financial datasets
Pattern Detection with Unsupervised Methods
This programme applies clustering and dimensionality reduction to identify recurring structures in cryptocurrency data - market regimes, correlated asset behaviour, and anomaly periods. No labelled data required.
- K-means and DBSCAN on market microstructure
- PCA applied to multi-asset correlation matrices
- Visualising cluster boundaries across time
- Practical use cases for regime detection