Marvik

The first project involved Marvik working with GU Trade, a company that provides visibility of execution at Point of Sale for mass consumption and distribution companies. GU Trade approached Marvik with the goal of automating the physical counting process of products across different brands. Marvik developed a system using computer vision techniques to determine the share of a given brand’s shelf at supermarkets based on a picture or set of pictures. The project was initially set for 3 months but was extended as new outcomes were identified. The project involved the CTO and a Delivery Manager, plus a dev team that changed based on the required skills. Technologies used included Python, OpenCV, Tensorflow, Keras, and Pytorch, and the system was deployed to Azure.

The second project saw Marvik working with a UK-based SaaS company that creates aging and lifestyle effects to show how a person could look depending on different lifestyle habits. The company needed Marvik to help find a solution for some of their outdated aging effects and implement neural and generative adversarial networks in their service. Marvik implemented a series of networking services for the aging effects, aiming to find a solution that balanced quality, cost, and speed. The project began in January 2020 and is ongoing.

The third project was with Resilience International Lab, an organization focused on addressing issues of climate change through strategic planning, technical planning, and predictive analysis. The goal of the project was to generate insights about how to deal with time series corresponding to different geographic points, using the data collected from monitoring stations for predictive analysis and interpolation. Marvik performed an Exploratory Data Analysis to understand the data availability and quality, then proceeded to train machine learning models and develop a proof of concept to predict water quality components. The project involved Marvik's co-founder and the CTO.