Marshall, co-founder of Juno, joins FLARE Perspectives to unpack how the AI-powered chronic-care platform grew to around 80,000 users in six months, why more than 1,000 conversations with users have shaped its product strategy, and what Juno has learned about turning deeply personal healthcare problems into products with real utility and remarkable organic growth.

We also explore the challenges of chronic-disease diagnosis, Juno’s long-term vision for patient and clinician support, and the optimism, resilience, and relentless learning required to build from scratch.

Will walks us through DIHI’s innovation pipeline—from identifying high-impact problems and evaluating proposals to piloting solutions across Duke Health. Through Sepsis Watch, he illustrates how thoughtfully implemented AI can support earlier intervention, improve clinical outcomes, and serve as a second pair of eyes for care teams.

We also explore AI’s potential to reduce administrative burden, return clinicians’ attention to patients, and create new opportunities for students interested in healthcare innovation.

Highlights

• Turning ideas into clinical impact: How DIHI selects, pilots, and scales innovations inside a working health system.
• AI as a clinical support system: Sepsis Watch helps care teams identify at-risk patients and intervene earlier while keeping clinicians at the center of decision-making.
• Measuring what matters: Successful healthcare innovation must demonstrate tangible improvements in clinical workflows, compliance, and patient outcomes.
• Returning attention to the patient: AI can reduce documentation and administrative burdens that contribute to clinician burnout.
• Breaking into healthcare innovation: Curiosity, adaptability, and firsthand exposure to real clinical problems matter more than arriving with every technical skill already mastered.

Natalie explains how DiagnoBac is rethinking biosensor development by treating the molecule, sensing hardware, biological environment, and computational model as one integrated system. We explore why experimental feedback loops and proprietary performance data—not foundation models alone—may become the real competitive advantage in AI-enabled biotechnology, as well as the challenge of understanding not only whether a molecule works, but why.

She also shares her vision for greater biological data standardization and collaboration, and reflects on the resilience, adaptability, and willingness to keep learning required to build a deep-tech venture.

Episode highlights

• Biosensing is a system-level problem: Binding affinity alone is insufficient. Molecules must be designed around the materials, biological matrix, hardware, and operating environment in which they will function.
• Prediction is not mechanistic understanding: Models may identify patterns that predict success or failure without revealing the biological mechanisms responsible—limiting our ability to explain and systematically improve performance.
• Data fragmentation slows the field: Greater standardization and responsible data sharing could create richer datasets and accelerate progress across AI-driven biology.
• Criticism is information: Natalie views rejection and critical feedback as signals that help founders refine, reposition, or pivot their ventures.
• Founders do not need to know everything: What matters is the confidence and openness to learn quickly, adapt continually, and solve unfamiliar problems as they arise.

Drawing on more than two decades across semiconductor R&D, manufacturing, product management, and business development, Chris explains the difficult transition from invention to commercially viable innovation.

We explore why VioNano is pursuing a customer-driven product model, how the AI boom is reshaping semiconductor manufacturing, and why even nanometer-scale improvements can create enormous economic value.

Chris also shares lessons for aspiring deep-tech founders: understand the customer’s pain point, gain industry exposure, build a trusted network, and be ready when the market finally catches up with the technology.

In this episode of FLARE Perspectives, the CEO & Co-Founder of 83 Sciences shared how that realization shaped the company he is building and ultimately pulled him from industry into entrepreneurship.

The conversation went beyond the familiar claim that “AI needs better data.”

We explored:

• What “better data” actually means in R&D when critical context is scattered across workflows, records, and individual teams
• Why so much experimental knowledge remains difficult for AI systems to use
• Where lasting advantage comes from as foundation models become increasingly accessible
• How proprietary experimental knowledge can become a compounding advantage
• What Ian learned moving from industry insider to founder and building 83 Sciences around a problem he had experienced firsthand

A founder conversation about AI-driven R&D, scientific data, company building, and the infrastructure behind the next generation of research.