Monte Carlo simulations and factor modelling have been at the root of complex option pricing theory. This competition invites machine learning enthusiasts to find the non-linear representation between model inputs and outcome in order to create a faster pricing engine without compromising accuracy.

About the sponsor

UBS AG is a Swiss multinational investment bank and financial services company founded and based in Switzerland. Co-headquartered in the cities of Zürich and Basel, it maintains a presence in all major financial centres as the largest Swiss banking institution in the world.



Become a financial engineer and solve a real life investment product problem. Use the dataset of factors identified by UBS quant team to train and validate your machine learning model. Your model needs to learn from the data and be able to predict the price of an equity structured product faster than traditional computationally intensive models. Efficiency is key to optimizing sales operations and the ability to generate a quotation faster provides a competitive edge. Solve efficiently the pricing puzzle with machine learning to extract this edge.


This is a supervised regression problem. Your target is to predict the value of the "val_lvsvcharge" column.

Coding environment

This competition requires participants to use Python. All data and coding environment are provided by Alphien. From the alphien dashboard, access IDEs, Notebooks and the Quantitative library centre on under the Research menu on the left. Participants cannot download the dataset to work locally. All research work has to be done on the platform directly.

Selection criteria

It is a pre-requisite that models have a high explanatory power, i.e. that predictions can be traced. Participants need to keep this in mind if they decide to use neural networks for instance: they have to design a property to trace how a prediction is made. Conversely, this property already exists for tree-based models.

The models will be ranked based on the following criteria: The Maximum absolute error (50%), Mean Square Error (30%) and the model Explainability (20%).

Among the models that can be considered, rankings will be established based on a the test set.


This competition is part of the UBS Hackathon qualification round. Please refer to the main page for award details.


UBS quant team has selected a curated set of factors to price structured products.
The dataset consists in around 14.5 millions by 165 data points, all formatted as floating numbers. Uncompressed, the dataset weights around 30 Go, making it challenging to work with. Consequently, the dataset has been split down and compressed.

Use the tutorial notebook to see how to retrieve the dataset for this competition. The tutorial notebook can be found in the Dashboard section


Support tools

  • After you have joined the competition, click the "get started" button. You will be redirected to your dashboard. In "Public Notebooks", filter notebooks by competition tag using the "UBS pricing" tag.
  • Make sure you have trained and validated your model to ensure generalisation and maximise your chances to win.
  • Use the forum to find help from the community or from Alphien support team.

Tips and advice

  • Use AlphienLab to conduct your research.
  • Once you have fine-tuned your model, document it in a notebook and submit the notebook for our review.
  • We highly recommend you use the forum as much as you can to find help.


Full competition rules can be found here.

For other rules details, please review the legal section at the bottom of the page.


Please refer to Alphien Terms of Use.