Tech Veteran Chris Dwan Challenges Life Sciences to Build Infrastructure, Businesses Strategically

By Allison Proffitt 

April 23, 2025 | After years of contributing to the Bio-IT World community as a speaker, awards program judge, and leading consultant, technology expert Chris Dwan took the plenary stage at the 2025 Bio-IT World Conference & Expo, sharing wisdom gleaned from over two decades of building computing infrastructures for the New York Genome Center, Broad Institute, and Sema4 and others. Dwan offered strategic principles for life sciences technology leaders that transcend rapidly changing technical landscapes and challenges for the culture of our space.  

"Technology is not ‘why’. It's not mission. Technology is not even ‘what’. What are you trying to accomplish? I'm your how-to-do-it person," Dwan explained, emphasizing that technologists must understand the scientific mission they support or risk merely "hobby shopping" with the latest trends. 

Dwan walked the audience through the evolution of high-performance computing environments he's developed since 2002, demonstrating how computing power that once filled rooms now occupies fractions of standard server racks. Despite these dramatic advances, he recommended looking for the principles that do not change over time and building your compute strategy on top of those. Dwan identified enduring physical laws around power, cooling, and information transfer that remain relevant regardless of technology cycles. 

Computing architecture belongs next to cheap renewable power, someplace where it's straightforward to cool, and close to high-quality fiber, Dwan advised, pushing back against recent claims that traditional data centers are obsolete in the AI era. “It turns out that despite what people will tell you, all systems are a little bit hybrid.”  

Addressing data management strategies for scientific research, Dwan highlighted The Cancer Genome Atlas’s Levels of Data. Raw and normalized data—levels 1 and 2—are immutable, he pointed out. “You can keep that forever; this guides your data strategy,” he said.  Beyond the raw and normalized data, though, all data are model-dependent. He stressed the importance of understanding the scientific models underlying data interpretation, cautioning that confusion often arises when terminology is overloaded. For example, What does “vector” mean to a mathematician, biologist, or epidemiologist? 

On business strategy, Dwan presented the classic "fast, cheap, good" trilemma, noting how biotech companies typically prioritize speed during early discovery, quality when approaching clinical development, and cost efficiency when reaching commercialization. He urged technology leaders to understand their organization's current priorities rather than attempting to optimize for all three simultaneously. 

Looking toward emerging technologies, Dwan predicted an "AI hangover" similar to the "cloud hangover" many organizations experienced after overcommitting to cloud infrastructure. "I think the AI hangover will be when we realize that domain expertise matters," he cautioned, suggesting organizations may regret dismissing experienced specialists in favor of AI-driven approaches. 

He further warned of the risks of not developing junior talent. Expertise is built by "the practice of the largest survivable failure," Dwan said. “It is terrifying!” But by allowing early-career team members to learn through challenging assignments rather than replacing their development opportunities with automated solutions. "We cannot just use the technology of the day to summarize or read or propose an algorithm or propose a solution. We have to keep developing our junior people. It is the only way we get experts," he insisted.