Devices used to detect and prevent wildfires in remote forests are expensive, but the one that Northeastern University student Anson He is making will be cheaper to launch on drones over dense woodlands. He is pursuing his master’s degree in computer science at Northeastern’s Vancouver campus. In January, he started a co-op at Bayes Studio — a Vancouver company that uses robotics and machine learning to make forest fire detection tools. He is helping to produce a device that uses less expensive components than others on the market. His role is core to the small company’s success: He is in charge of prototyping the hardware and coding the software for what Bayes calls its Edge device.  Other team members work on integrating artificial intelligence into the device’s functionality and connecting the device to servers.