Episode Description
In this episode of Neural Compass, host Mark Jacobstein (co-founder and president of Jimini Health) speaks with Shreya Rajpal, founder and CEO of Guardrails AI—an open-source pioneer at the forefront of AI safety and infrastructure.
They explore what it really means to make AI reliable, diving into lessons from self-driving cars, the limitations of traditional “accuracy” metrics, and how to build systems that earn human trust. From simulation testing to open-source strategy, Shreya offers a compelling look at what it takes to scale AI safely—and why responsibility starts long before deployment.
This episode touches on:
- Why reliability is a lifecycle challenge, not just a technical one
- How simulation testing is reshaping safety in GenAI
- Why accuracy isn’t the right goal—and what to measure instead
- How product design and trust go hand in hand
- The engineering mindset behind building a company
About the Speakers
Shreya Rajpal, Guest
is the CEO and founder of Guardrails AI, an open-source platform for ensuring trust and reliability in generative AI applications. Previously, she worked on machine learning for self-driving systems and special projects at Apple. Today, she’s one of the leading voices shaping AI safety and developer experience, with a mission to make scalable, responsible AI the default—not the exception.
Mark Jacobstein is the co-founder and president of Jimini Health. A longtime entrepreneur at the intersection of AI and health tech, he now leads Jimini’s mission to transform mental and behavioral health through clinically grounded, AI-powered care.
For healthcare executives, researchers, investors, and anyone who thinks the mental health crisis deserves serious answers — this is the show.