Brave Innovation

Synthesis speaks with David Ford, Senior Director, Digital Technologies at Cambrex, to find out what the intersection of complex chemistry and lab innovation means for the pharmaceutical industry moving forward.

What sparked your interest in science and technology?

As a kid, science and technology were ever-present. My dad is a biomedical engineer and my mom is a chemical engineer. My mom ran her own filter testing company, where I worked when I was young; I started out on garden duty, picking up rocks to save the lawn mower, before graduating to the lab environment when I was a little older.

At college (UC Berkley), I followed my mom’s footsteps into chemical engineering but discovered that the forward-looking research topics pursued by the faculty (futuristic microfluidics, biological applications) were not fully reflected in the undergraduate curriculum. Much of the advanced coursework was geared to a career in the petrochemical industry, which wasn’t for me.

I felt that chemistry was a better fit, so I made the switch. I was attracted to chemical biology at first, but ended up falling in love with organic chemistry. Completing my PhD studies at Harvard University with Eric Jacobsen was a great experience. What sticks in my mind the most is the amount of freedom that I was given to pursue different research avenues. If a project took a two-week detour to try something totally speculative, I didn’t have to convince anybody. And that was great for somebody like me, with a lot of curiosity and a tendency to tinker.

Your first job in process development must have been quite a shift in gear…

It was certainly a big difference from grad school! I was still doing organic chemistry, but with a flavor of chemical engineering. I had to quickly learn how to set realistic goals and commit to completing tasks on time because I was part of a team, with people depending on me.

After a few years, I was drawn to the culture of innovation at Snapdragon Chemistry (now a Cambrex company). Before I joined in 2017, I visited Snapdragon in a different context and remember watching a trained chemist (who is still on the team) soldering LED chips onto a circuit board to build a prototype photoreactor. Her plan for the day was to build a photoreactor to run a novel reaction. I really respected the bravery and was inspired by the can-do attitude: “I’m going to go buy some chips and a power supply, grab a soldering iron, watch some YouTube videos and just go for it!”

That culture of brave innovation defined the way Snapdragon Chemistry was then and still is today.

What’s the biggest change or advance you’ve seen in the last decade?

Digital technology has been entering process development labs for the last 20 years or more. But in the last decade specifically, I’d say the biggest change has been putting innovation and laboratory automation directly in the hands of scientists and engineers who are not automation specialists.

We’re increasingly seeing a focus on laboratory automation, interest in closed loop optimization, and the urge to extract full value from technology. And that trend has accelerated even more dramatically in the last 12 months thanks to the vast improvement of AI tools.

The power of digital tools and AI brings us to your new role as Senior Director, Digital Technologies…

The idea of digital transformation and building digital business processes is not new, of course. The need to fully digitize information on paper or “paper on glass” (basic spreadsheets, for example) is well recognized. After all, without structured, accessible data, you have very limited ability to connect processes together, automate them and gain insight at scale.

With AI, the cost (measured in time or dollars) of digitizing a process is much less than it was 12 months ago. At the same time, if you have a digital process, you’re much more enabled to use AI to interrogate it.

In the early days of Snapdragon, we were already actively working in this space. Partly because of the culture of innovation, but also out of necessity. We were facing complex chemistry challenges, but we didn’t have the luxury of spending “big tech” budgets. Perhaps we were lucky to have the right innovative people at the right time, but we quickly found out that it wasn’t as hard as we feared to build our own hardware and software tools. Any sense of trepidation was quickly replaced by excitement about our new-found ability to build tools that were tailor-made for the tasks that we had in the lab.

The main software tool we were developing back then and are still developing now is LabOS™. Put simply, it’s a tool that allows us to collect data from all the equipment in the laboratory, while also acting as an electronic lab notebook. LabOS has evolved a great deal over the intervening decade, but our excitement for the power and flexibility it provides has not waned.

Today, we can build useful tools extremely quickly and on an even larger scale than ever before. I personally find it really rewarding to see different processes connect with each other to deliver that networked effect.

What does your personal focus and Cambrex’s evident investment in digital technologies mean for partners?

Disconnected processes commonly involve a lot of repeated work, where you have to go back to source documents and/or re-enter data in a new format. It’s a waste of resources and it’s not fun, either.

By digitizing processes and connecting them, we can build one digital model of the entire execution of a project or a process as it develops, which allows anybody who interacts with the project to more rapidly retrieve more intelligent answers. As the project advances, we can more quickly and more faithfully translate processes into the R&D lab, through different engineering and safety studies, into manufacturing and then on to continuous improvement – all while taking advantage of the work that was done before.

Crucially, with all those digital data streams connected and the ability to use AI/ML to interrogate the resulting information, we can also gain greater process understanding in a much shorter time frame.

Where have you already seen the benefit of the close interplay between complex (process) chemistry and digital technologies?

Initially, we started building digital tools in the context of continuous manufacturing and flow chemistry. Back then, continuous manufacturing was considered something radical that was only used in the most desperate situations in chemistry; for example, cases where a company had already tried to scale up a process only to be faced with a batch failure.

With such inherent complexity, we knew we needed to be able to interrogate data to explain issues that would otherwise be complete mysteries. Rich data gave us insight, direction and confidence to make the right changes for good reason.

Over time, our understanding of continuous manufacturing improved – thanks to that all-important data – and we found ourselves being able to apply the approach in cases that were less dire.

The field of peptide synthesis is another area where digital technologies have been really impactful – again born out of necessity (in this case, the need to meet the demand driven by exploding interest in peptide drugs, such as GLP-1 receptor agonists). Liquid phase peptide synthesis is an attractive alternative to solid phase synthesis, but it demands many liquid–liquid extractions, which can be difficult to automate and laborious to do by hand. Designing and building digital tools to control those processes, while retaining both flexibility and reliability, has made a huge difference to how we actually execute the science.

Though desperation provided us with the initial motivation to learn how to perfect these tools, our successes have lowered the barrier for future chemists and opened doors to alternative paths forward.

An impossible question, but how do you think the next decade will play out?

You’re right; that’s an impossible question, but in light of recent advances in AI, it is one that people are asking themselves across every industry and every facet of life.

I think it’s hard to overstate just how dramatic a change we’ve witnessed over the last 12 months. Yes, the tools were already available 12 months ago (and some forward-looking colleagues of mine were using such tools 2-3 years ago), but the results weren’t as impressive. Whether you used AI assistance or not, the overall time to complete a task wasn’t dramatically different.

Today, we are seeing a five- to ten-fold acceleration of software development tasks. The speed at which we can build new digital processes and interconnect them is faster than it has ever been. Approaches such as high-throughput experimentation and closed loop optimization have, until recently, existed on the periphery – often only considered when facing particularly challenging situations. I predict that we’re going to see such approaches really stepping into a central role in day-to-day work.

We are increasingly being given the opportunity to stay laser-focused on the most important aspects of our objectives and granted new insight into how they fit into the big picture. We are able to benefit from past experience by accessing information that is more readily at our fingertips. We can rapidly consider how different parameters might impact a process at a different scale. We can more accurately find answers to questions that may have been economically infeasible to answer. And we are increasingly (and thankfully) quickly shifting from best guesses to informed decisions.

Science is very exciting. But it can also be repetitive, challenging and frustrating at times. Beyond a quicker path to the answer or even success, I believe the right digital technologies and tools will also make the work more rewarding and fun.

What excites you most about the future?

What excites me most is that we have not yet reached equilibrium with AI. And that means there is a huge opportunity to explore the value of building and connecting smarter tools that have simply not existed before. The possibilities are nearly boundless.