According to GeekWire, a Seattle-area startup called Pauling.AI, founded in 2024 by former Google technical director Javier Tordable, is using artificial intelligence to automate early drug discovery. The company’s platform performs computational chemistry work, engineering drug candidates and modeling their interactions, tasks that used to take three to six months but now reportedly take just weeks. The startup, which has six remote employees and undisclosed pre-seed funding from Flex Capital, operates on a “scientist-as-a-service” model, serving less than a dozen customers including high-profile academic institutions. Chief Scientific Officer Oleksandr Savytskyi, a computational biologist with experience at the Mayo Clinic, leads the scientific efforts. Tordable’s ultimate vision is to help increase new drug approvals from 30-40 per year to 300-400, specifically to tackle rare diseases big pharma often ignores.
The AI Drug Rush Is On
Here’s the thing: Pauling.AI is jumping into a field that’s getting seriously crowded, especially in the Pacific Northwest. The article name-drops competitors like Vancouver’s Variational AI, Seattle’s own Potato and Synthesize Bio, and the well-funded Xaira Therapeutics, which has labs in Seattle. That’s a lot of smart people and a ton of venture capital all chasing the same holy grail: using AI to brute-force the discovery of new molecules. It feels like the early days of self-driving cars, where everyone had a different technical approach but the same grand promise. The question isn’t really if AI will change drug discovery—it’s which of these approaches, or which combination of them, will actually work consistently and get a drug across the FDA finish line. Pauling’s “scientist-as-a-service” angle is interesting because it’s essentially offering AI as a tool for existing researchers, rather than trying to be a full-stack biotech that runs trials itself. That could be a smarter, capital-efficient way to start.
The Google Brain Drain Effect
I think the founder’s background is one of the most telling details. Javier Tordable spent 16 years at Google, most recently directing healthcare and life sciences initiatives. And he openly admits he’s not a biologist or chemist. That’s the story of this wave of AI biotech in a nutshell: it’s being built by brilliant engineers and data scientists from Big Tech who believe complex systems can be modeled and optimized with software. They’re looking at biology as an information processing problem. This has huge potential, but it also introduces a cultural gap. Drug discovery is famously brutal, iterative, and full of biological surprises that don’t care about your elegant algorithm. The success of these ventures, including Pauling.AI, might hinge on how well that Google-style engineering mindset integrates with the deep, messy biological expertise of people like Chief Scientific Officer Oleksandr Savytskyi. If they can truly speak each other’s language, that’s when the magic might happen.
The Real Prize: Rare Diseases?
Tordable’s stated mission to go after rare diseases is noble, and it’s the kind of thing every startup in this space says. It makes perfect sense as a goal. If you can drastically shrink the time and cost of the initial discovery and design phase, you theoretically make it profitable to develop drugs for tiny patient populations. But let’s be a bit skeptical for a second. The economic model still has to close. The later stages of development—clinical trials, manufacturing, regulatory approval—are astronomically expensive and time-consuming, and AI doesn’t really touch those yet. The promise is that by creating better, more targeted drug candidates from the start, you increase the odds of success later, saving billions in failed trials. That’s the exponential return. If Pauling and its peers can prove they consistently produce candidates that are more likely to work, they won’t just be serving academia; every big pharma company will be knocking on their door. The potential benefit to humanity is enormous, as Tordable says. But first, they have to prove the model works at scale. It’s a huge bet, but one absolutely worth taking.
Beyond The Hype
So what does this all mean for the actual pace of new drugs? Proponents talk about a 10x increase, from 40 to 400 new approvals a year. That’s a staggering thought. It would revolutionize medicine. But we’re not there yet. These are still the early innings. The field is fragmented, and for every AI-discovered compound that enters the clinic, there will be many that fail—because biology is hard. The nonprofit FutureHouse, mentioned in the article, is a good reminder that not all the work is happening in venture-backed startups. The real impact in the near term might be more subtle: giving academic and small biotech researchers powerful, on-demand tools they could never afford or build themselves. That democratization could indeed unlock novel science in overlooked areas. The timeline from “AI-designed molecule” to “medicine in a vial” is still measured in years, not weeks. But if companies like Pauling.AI can reliably cut months off the front end of that journey, they will have earned their place in the story. Now we wait and see if the data holds up.
