Case Study: Riskfuel slashes compute costs and speeds financial risk modeling with Weights & Biases

Riskfuel slashes compute costs and tracks 20+ ML projects with Weights & Biases

Riskfuel, a Toronto-based AI company for capital markets, faced the challenge of helping financial institutions reduce the enormous compute costs and time required to run intensive Monte Carlo simulations for pricing complex derivatives. Their goal was to provide real-time risk metrics without sacrificing accuracy. They partnered with the vendor Weights & Biases and used its experiment tracking platform and W&B Models to manage their complex machine learning workflow.

The solution implemented by Weights & Biases provided Riskfuel with a robust and scalable way to track multiple ML projects, monitor model performance, and identify error patterns. This allowed Riskfuel to successfully build deep neural network replicas of their clients' pricing engines, which execute much faster. The results included significantly slashing compute costs for their clients, enabling real-time risk calculations, and providing a major environmental impact by reducing total compute required. Weights & Biases was crucial in helping Riskfuel manage their intricate development process and focus on their core mission.

Riskfuel