Multiply Labs Uses AWS RoboMaker to Manufacture Individualized Medicines

Personalization is the future of medicine. Currently, daily drug regimens can be challenging and complex, with patients taking multiple prescription drugs to treat different medical issues. Multiply Labs strives to solve this problem by manufacturing a single daily capsule containing a patient’s entire prescription—with a dosage individualized to a patient’s needs. The company supplies robotic manufacturing-as-a-service for this customized medicine production. “Medicine is becoming more personalized, and we help companies enable that by using robots to create pharmaceuticals in individually sized batches,” says Zack Bright, chief technology officer for Multiply Labs.

Because its software must comply with strict Good Manufacturing Practice (GMP) regulations from the U.S. Food and Drug Administration (FDA), Multiply Labs needed to securely and automatically deploy software to the robots filling pills. “The software configurations for our robotic manufacturing cells must be monitored very closely so we can ensure traceability for compliance reporting,” Bright says. “However, we relied on inefficient processes that included manually filling out stacks of paperwork for each batch. We needed a more automated process if we were going to grow our business.”

Multiply Labs decided to move its robotic software environment to Amazon Web Services (AWS) and take advantage of AWS RoboMaker to automate deployment of updated code to its robot application, based on Robot Operating System (version ROS 2). “AWS RoboMaker was ideal for us because now we can manage nine robot systems in parallel from our quality control software,” says Bright. Previously, the organization relied on insecure SSH tunnels and scripts to pull and compile code programmatically—a complex process that, he says, “wasted engineering time and talent.”

Using AWS RoboMaker, Multiply Labs deploys its robotic application multiple times a day to its robot fleet. “We hadn’t done this type of testing before,” Bright says. “We ran smoke tests and some regression tests, which took weeks.” Multiply Labs also plans to incorporate simulation testing of its robots in the pipeline. “Without simulation testing, we would have to shut down the manufacturing line to test. Simulations mean we can maximize uptime as a robotics-as-a-service company.”

In addition to implementing AWS RoboMaker with its continuous integration pipeline, Multiply Labs also uses Amazon CloudWatch to log all versions of each drug manufacturing batch.

The company uses AWS CodeDeploy and AWS CodePipeline to set up and deploy software builds, and it relies on Amazon DynamoDB to store data. Its robots interact with servers that send ROS 2 batch data in and out of an Amazon DynamoDB database. “By storing the batch records digitally on Amazon DynamoDB, we have a repeatable and traceable way to store data,” Bright says. “We can clearly show that an exact amount of a patient’s medication was filled by a certain machine in a certain order.”

With a GMP integration of Amazon DynamoDB and ROS 2, plus AWS RoboMaker, Multiply Labs can enable digital batch records to be made by robots. “Before, our paper records were hundreds of pages long and took hours to fill out and review by hand,” says Bright. “We were lucky to handle two batches at a time. Our new stack enables dozens or, as we scale, hundreds of batches to be manufactured at once. The flexible storage schema in Amazon DynamoDB, with a rapidly scalable capacity, means we can handle this for all our customers and store and manage thousands of batches at once at full scale.”

Using AWS, Multiply Labs can enable consistent, controlled software deployments with a validated image every time. “We can ensure the software used during manufacturing was deployed exactly per our gold standard by using AWS RoboMaker and AWS CodePipeline,” says Bright. “We automatically load the software onto our robots, multiple times a day, immediately before the robot runtime. There’s no chance the wrong version of the software is running or that anything could be tampered with. Because of AWS RoboMaker, we can achieve continuous compliance for our FDA-regulated workloads and demonstrate traceability.”

Multiply Labs delivers personalized medicine more easily by relying on AWS. “The best way to do so is by using robots to make customized pharmaceuticals in individualized batches,” says Bright. “That requires a fully automated and parallelized system where robots are in perfect sync. Using AWS RoboMaker, we can centrally orchestrate each manufacturing cluster of robots, with 9–15 different robots all connected to AWS RoboMaker. Each robot has different tasks such as transporting, weighing, or sorting capsules and then packing and sealing them. As a result, we can deliver customized therapies to patients in a single pill. We are also ready to scale this to all our customers quickly. Using AWS RoboMaker, it will be easy for us to go from our first nine robots to our next 100.”

Using AWS RoboMaker to automate deployment, Multiply Labs can track, trace, and inspect robot software to a level that would normally require time-consuming work. “Creating a solution like this on our own would have taken us a year of development work,” says Bright. “It would not have been possible to do what we wanted to do without a tool like AWS RoboMaker.” Deployment is automated and integrated into the company’s continuous integration and continuous delivery (CI/CD) pipeline, so engineers can focus on robot programming, not deploying code. “This easily saves an hour or two a day for our engineers. Over the next year, this would effectively be tens of thousands of dollars saved.”

Multiply Labs will also meet its goal of providing a manufacturing-as-a-service model to pharmaceutical companies. “Having an AWS-based system allows us to deploy both hardware and software in a scalable way, so we can sell our manufacturing capacity as a service,” Bright says. “And we can deploy repeatedly, using separate instances to assure our customers that their data is protected and not seen by others.”

With these capabilities, Multiply Labs can support customers’ clinical trials. “We can repeat instances of our software, put it into a shipping container, and send it directly to a company starting a clinical trial,” says Bright. “It effectively looks the same to our customers who deploy our systems abroad or have us manage the robots in our facility. We can manage the software from our San Francisco headquarters, no matter where the trial is across the globe. Meanwhile, our customers can see that the batches are being fulfilled in real time whether the robots are on-site or in Multiply Labs’ facilities. Using AWS definitely helps us meet our strategic goals.”