Articles
Building a Forecasting System (from Scratch)
Let’s address the elephant in the room: there is a significant gap in forecasting between research and its practical application in industry. This article documents the design and engineering of an end-to-end Demand Forecasting System.
“In-Context Learning”, Synthetic Data and Evaluation Issues
Entering 2026, we are now roughly two and a half years into the era of Time Series Foundation Models (TSFM). This article discusses the nature of ICL in forecasting, the role of synthetic data, and the evaluation crisis in the field.
Are Transformers Effective for Time Series Forecasting?
For the past year, I’ve been asking myself the same question: “Are Transformers Effective for Time Series Forecasting?” Recently, I came across this paper that discusses this topic in more detail, exploring the limitations of Vanilla Transformers.
Forecasting with Multiple Temporal Aggregations. Part 1: Uncertainty and Forecast Combinations
This series of articles is about time series forecasting using Multiple Temporal Aggregations (MTA). In this first part, I give a brief introduction to the motivation behind forecasting with MTAs and the benefits of Forecast Combinations.
Forecasting with Multiple Temporal Aggregations. Part 2: Introduction to Temporal Aggregations and … MAPA.
In the second part of this series, I introduce MTAs and talk about the first forecasting method that effectively applied them in practice, focusing on the intuition behind temporal aggregations and MAPA.
