4.12.22-Health-Science-Startups-Multiply-Labs

FRED PARIETTI: Hi. I'm Fred, co-founder of Multiply Labs. And it's great to be back here in Cambridge. I got my PhD here in MIT, 2017. And, well, actually, pretty much 80% of our company-- it's either MIT mechanical engineers or MIT computer scientists. And so you might wonder, what are all these robotics people doing in the pharmaceutical space? And, well, we strongly believe that robotics is the only way to scale the manufacturing of next-generation therapies and, in particular, cell therapies.

So what's the state of the art today? Today, cell therapy manufacturing systems are extremely serial. So you see on the left, essentially, this is state of the art today, really. It's a set of single-unit operation machines that are individually functionally closed. But the transfer of the product from one system to the next-- it's done manually. And so the bottleneck, clearly, here is the manual labor, which is highly skilled, hard to train, hard to retain, that's needed to run this process.

Some system have brought all of these unit operation into the single machine. But this machine is serial, which means that at any given time, only one subsystem is active. So it's structurally underutilized and, as you can see, still pretty manual.

So why is serial manufacturing wrong, basically, for cell therapy? Well, this system is not balanced. Like any manufacturing engineer would tell you, if something takes so much-- one or two orders of magnitude more than all the other processes-- for example, expansion, the bioreactors-- then that system is not balanced. And so having a serial system means that all the other machines are waiting for the bioreactor-- for the cells to expand.

And so what's the solution? Like I mentioned before, we strongly believe the only way to solve this is through robotics. And so here you can see the robotic systems that we developed. Well, in this particular case, these are used for small molecule manufacturing. But the concept is very similar. It's a set of modules around the robotic arm. These are already being deployed, by the way, in a GMP facility. And every single module performs a single task of the process. So we can use multiple modules in parallel to address bottlenecks.

All the planning is done in the cloud. But most importantly, all the recording process measurements, all the data-- they go back to the cloud in digital records. And so they can be integrated with MES systems. And so this level of automation really enables that level of scale.

So there are four key advantages of this approach. And I'm going to walk you through them. First one-- we work with state-of-the-art, with market-leading instrument makers. So we work with proven instruments that already have a long track record in GMP. And we put each of them into a separate module. So we, in fact, teach the robots to automate proven systems so that working with this kind of automation, you don't need to modify the process. You just select the modules that realize your process.

The second advantage is on the quality side. Right now, sometimes, we still see paper-based batch records. It doesn't matter if they are digital if they're still human input. The only way to scale is to have validated machines collect in-process data. If there is a single person, that's going to be the bottleneck. And so here you see in this other video-- is the same video as before. But now you see on the left a digital record that's been filled out in real time by the system. And it's in the cloud. It's accessible from anywhere. It's shareable with customers, with regulators-- so really avoiding the problem of human bottlenecks and human errors.

Of course, robots are very dense. I don't need to say this. We can pack them. We can stack them. The level of [AUDIO OUT] is unprecedented.

And of course, parallelism-- that's very important for efficiency. One bioreactor means everything else would wait for it. Multiple bioreactors in parallel to the tune of hundreds, really, is what enables these to scale. We're not talking about making 100, 1,000 cell therapies. How about making 100,000? And parallelism and robotics are really the answer here.

So partners-- we already work with Cytiva, which is actually an ILP member. They are manufacture, of course, of leading bioreactors. We're about to announce several additional partners that are still confidential. And we are always interested in looking for market-leading GMP instrument makers to integrate them into our modular system. Thank you very much.

[APPLAUSE]