How Faro Uses AI w/ CTO Patrick Leung

Patrick Leung is the CTO of Faro - a software that helps pharma giants design multi-billion dollar clinical trials.

His team is using AI to move faster than ever before, inside one of the most complex, compliance-heavy industries out there.

We spoke to Patrick about where AI is most useful in the dev cycle, why code review is becoming the new bottleneck, and how Faro saves pharma companies over $100M per trial.

Catch the full conversation on YouTube, or read our takeaways below:

Shipping fast inside a regulated world

Pharma isn’t exactly known for speed. Clinical trials can last a decade and cost over a billion dollars. Every new change needs review boards, compliance sign-offs, and endless documentation.

When Patrick joined Faro, he was stunned to find how clinical trials were designed.

“Believe it or not,” he said, “these multi-billion-dollar projects are designed in Microsoft Word.”

How do you meaningfully leverage AI in such a regulated world?

Prototyping at the speed of thought

Patrick’s team uses AI to prototype, test, and iterate faster. Before AI, building a proof of concept could take weeks. Now, they can spin up a near-production prototype in hours.

“Almost nobody is Steve Jobs,” Patrick said. “We all need to stumble around and explore the solution space through experimentation. And now, AI makes that experimentation fast.”

Figma mockups have given way to working code prototypes, built directly with AI tools like Cursor and Claude Sonnet. Even adjacent teams (design, QA, DevOps for example) now use AI to automate testing, configuration, and internal docs.

The new bottleneck: code review

With developers coding twice as fast, the bottleneck has now shifted to code review.

“AI’s made engineers dramatically more productive... but that means twice the code to review, with the same number of reviewers.”

Faro’s already exploring AI-assisted code review tools that can flag design inconsistencies and enforce internal idioms. Patrick thinks the next velocity unlock will be reliable systems that make sure AI-generated code still meets the same standard of craftsmanship.

AI across every adjacent function

Faro’s AI adoption goes far beyond engineering. UX, QA, and DevOps teams are automating the tedious layers of their work, from test generation to infrastructure as code. Even marketing uses AI to draft technical copy and customer literature.

“Every team can use it. If there’s a pattern, AI can automate it.”

Hiring for curiosity, not just stack

Patrick has also changed his hiring philosophy. Building with AI requires a new kind of engineer: one who can think in systems, not scripts.

“Using AI isn’t as simple as talking to ChatGPT. It’s complex. You need people who’ve actually built systems with it.”

Like many of the other leaders we've spoke to for this series, Patrick’s hiring playbook now prioritizes curiosity and adaptability over specific frameworks

In Patrick’s words:

“AI won’t replace people. But people who can effectively use AI will replace people who don’t.”

Actionable takeaways for builders

• Prototype early. Use AI tools like Cursor to validate ideas in hours, not weeks.
• Focus on code review. AI sufficiently takes care of code generation. Code review and testing will need automation next.
• Apply AI across functions. Look beyond dev. Automate design, QA, and ops.
• Fix at the design stage, save big. The ROI of AI is highest where rework costs the most.
• Hire for mindset. Curiosity and agency matter more than proficiency with the current trending tech stack.

This is not about implementing AI for AI's sake. Faro's shipping velocity directly impacts their customers. For instance, one of Faro’s customers, Merck, published a joint study showing AI-optimized trial designs can save over $100 million per trial. By running simulations early, teams can predict enrollment bottlenecks, reduce unnecessary patient assessments, and cut months of delay.

It’s the same principle Patrick applies to engineering: fix the design upfront, and you save exponentially later.

We had a great time jamming with Patrick! You can catch our full conversation on YouTube, alongside episodes with engineering and product leaders from Intercom, Monday.com, and Vercel.