While nobody can forecast the future, algorithms can definitely help and reduce uncertainty. Time-series forecasting has long been a personal interest of mine. I have also commonly seen and built solutions to time-series problems in industry. For example, forecasting of inventory levels is essential for ecommerce companies to manage cashflow, and modeling and decomposition of time-level data is essential for both understanding processes as well as identifying anomalies.
Coming from an experimental physics background building scientific instrumentation and dealing with noisy data acquisition, I am particularly interested in companies with remote sensor data, from industrial sensors to drone images.