Mothership

Our engineering team currently works on one-week sprints with weekly retro's. After careful planning and task allocation, engineers will generally work at their own pace throughout the week, leading to a peer review near the end of week or early the following week.

We are a sharp group of innovative engineers who are not afraid to approach problems with some out-of-the-box thinking. Because of this, we highly encourage group planning and architecture meetings where even the craziest of ideas are supported and evaluated. We strongly believe that to make a big impact you need to do things the next person wouldn't. We don't shy away from this challenge, we run towards it at full speed.

Our system automatically dispatches shipments to drivers all around the country based on a complicated machine learning system. The drivers' proximity to the shipment is only a small portion of the algorithm. The system takes into account historical driving patterns and preferences, traffic conditions, projected new shipments along the route... etc. Because of this, we have one of the hardest on-demand marketplace logistics problems to solve. It's not a one-job-one-driver system - one driver might be allocated to 5 or 6 different loads all of which have strict requirements of when they can be picked-up/dropped-off - think traveling salesman with dozens of node-specific variables.

We also integrate with dozens of legacy carrier API systems that require us to standardize I/O across the board to provide a seamless experience to the customer. These systems are often so rudimentary that we take it upon ourselves to build out extensive functionality on top of their existing system.

On the front end we are challenged with several complicated react application requiring performant real-time maps with hundreds or even sometimes thousands of markers. Traditional approaches would screech a map like this to sub 10fps rendering. However, our engineers have built out custom Web-GL layers boosting our render performance to 50+ fps even on the most demanding of maps.

Additionally, our systems demands hundreds of complicated sagas across dozens of store modules, not to mention most of them operating in real-time. For a front end react dev this would likely be one of the most robust, yet interesting systems to build upon.