PeopleTrack is our AWS-powered solution that offers retail store owners the ability to better understand customer behavior and engagement. This understanding can then unlock key decisions that promote better customer engagement, even seemingly small decisions such as better positioning of a certain product in the store. The auto-scaling features that AWS provides has enabled us to deliver our PeopleTrack to a continuously increasing number of retail stores nationwide while keeping our cloud usage costs in check.

Our VisionTrack solution leverages vision, real-time edge AI inference, and the cloud to eliminate low-skilled labor while performing at a high efficiency. We design our solution to be horizontally scalable, allowing us to provide solutions ranging from product defect detection all the way to automation of driving assessment. Instrumental AWS services such as Kinesis, EC2 and Sagemaker have enabled us to achieve requirements for real-time inference and streaming data in real-time and training models with blazing speed.

Deployed to over 500 lanes in Malaysia, our VehicleTrack solution is capable of making real-time inference at the edge to automate toll payments, producing key data such as images, vehicle classification and licence plate recognition that need to be streamed to the cloud. This huge volume of data then needs to be transformed and loaded into a database that enables powerful visualizations to be generated from those data. With AWS services such as IoT Greengrass, Kinesis, Lambda, OpenSearch, and Sagemaker, we are not only able to seamlessly deploy code from cloud to edge, but also to develop a data pipeline that allows for data-driven decision making. This subsequently contributes to faster iterations of model training that enables us to achieve model accuracy goals in a timely fashion.