In the world of electronic data, the cloud is the ultimate storage solution. Most products, though, can't be stored electronically. We can't digitize food, clothing, or even the phones and computers we use to access the cloud. All of these products take up physical space, which means they can only be stored in warehouses.
Google Cloud can't replace warehouses, but it can change the way they're organized for the better. Fulfilld is a warehouse management system using the Google Cloud Platform to ensure that the warehouses it services can manage their inventories as efficiently as possible. With hardware and software attributes such as microservice architecture and a variety of ML models and APIs, Fulfilld provides warehouses with management solutions that will make storage in those warehouses as functionally dynamic as storage on the cloud itself. C2C recently invited Fulfilld CTO and co-founder Michael Pytel to give a presentation and explain to our community what this looks like on the infrastructure level.
A warehouse is essentially just a large, open space, but once it's put to use for storage, it becomes much more. Products are shipped in and out, packaged and unpackaged, stacked and removed, and carried from place to place, sometimes by workers on foot, sometimes by large, fast-moving machines, like forklifts. Managing a warehouse requires navigating this space and anticipating this movement well enough to coordinate all this activity quickly and without incident. Pytel described how Fulfilld's ML models work in concert so that the app understands the warehouse environment intimately.
The Fulfilld warehouse scanner can collect location and position data from the surrounding area six times per second, (and in some scenarios, as Pytel revealed later, up to twenty-five times per second). This data collection can be used to identify storage space and to map routes to that storage space throughout the warehouse, so that transportation to that storage space is easily facilitated.
Here's Pytel explaining Fulfilld's location awareness capabilities below:
The Fulfilld scanning app also uses a 3D Bin Packing algorithm, coded in Python, to use the material dimensions of a specific unit to identify the best possible storage location for that unit:
Fulfilld's understanding of the warehouse environment also extends to the work that takes place within it. Fulfilld's location and position data make everything in the warehouse easy to find, as well as the best place to put it. The device also keeps track of the tasks that need to be accomplished in the warehouse, and which items and locations are associated with which tasks:
Breaking down the needs of the warehouse worker, from the location of a certain item, to the fastest route to that item, to the tasks the worker needs to complete associated with that item, reflects the microservices architecture at the structural core of Fulfilld's design. Pytel acknowledges that "every modern application today is built on a microservices architecture," but in Fulfilld's case, that microservices architecture is a natural component of an application run on Cloud Functions; as Pytel put it later during a question and answer session, "using Cloud Functions forces you to think in a microservices architecture." Here's a little more from Pytel on how that microservices architecture works for Fulfilld:
Pytel stressed repeatedly that the Google Cloud Platform and Google's ML models are what enable Fulfilld's sophisticated technology. Even working with Google's technology, though, involves negotiating between multiple options and weighing different considerations. Pytel outlined a few of these considerations, including the potential advantages of moving Fulfilld's runtime from Cloud Functions to Kubernetes, comparing and contrasting the different advantages of CloudSQL and Cloud Spanner, and looking at those different advantages alongside the unique advantages of Cloud Firestore:
These are some of the highlights of Pytel's presentation, but Fulfilld is a device with so many features and offerings that only the full presentation can do it justice. Watch the full video to learn more about what else Fulfilld can do, including advanced document scanning via language detection and image capturing, and a responsive digital assistant named Phil, who Pytel's co-founder says speaks in a Southside Chicago accent and who - - if Pytel's personal goal is met - - will be a better digital assistant than Siri.
Do you work for or with a warehouse? Do you have any warehouse management considerations you would want a device like Fulfilld to meet? Reach out and let us know!
