A major cause of customer interruptions (CI) in the power grid is due to vegetation growth. Proactive trimming can decrease CI’s, but can be wasteful if done without considerations given natural trimming due to weather.

An ensemble model was created to account for the seasonality, cycles and impact of weather events on CIs, and predict future CIs based on expected weather conditions. This prediction is then used as part of the trim plan.

The Levatas team worked with our client’s Power Delivery team to develop a computer vision solution to identify and catalog electrical components from drone aerial imagery.

This solution is used to drive more intelligent predictive maintenance to identify and repair faulty components avoiding more expensive equipment failures and/or costly outages.

A fleet truck company was exploring different options and ways to increase accuracy of pricing models with regards to vehicles. The goal was to optimize revenue annually and smooth the variance between projections and market values.

Levatas ultimately delivered was a production ready ensemble model that is part of an overall algorithm that can be used to better make decisions around selling trucks and tractors.

We worked with our client to build and deploy a model for identifying and counting boxes for inventory tracking in their warehouses.

The model was integrated with Levatas’ custom Spot Robot payload. The next phase of the project will integrated into their order management system for realtime ordering of items that are low.

Levatas developed a two-part solution for worldwide hospitality company. First, AI models were created to better serve the current recommendation engine for guest experiences.

Second, a chat function was created to integrate with the application, offering a robust NLP component which could run in an environment with limited computing resources, so that guests could naturally use voice and chat when seeking recommendations during their stay.

The Levatas team worked with our client to build a computer vision solution to improve guest experience. They use Computer Vision to count guests within their restaurants in order to better understand wait-times, inform staffing decisions and make better recommendations to guests.