AI-driven protein design is reshaping the global drug discovery paradigm. Since AlphaFold dramatically improved the accuracy of protein structure prediction in 2020, drug discovery has shifted from “moonshot” to practical strategy. The spotlight grew even brighter in 2024 when breakthroughs in protein structure prediction and design were recognized with the Nobel Prize in Chemistry.
Riding this wave of change, Galux has emerged as one of the most closely watched AI protein design companies in Korea and beyond. The company recently drew significant attention for its success in de novo antibody design across multiple targets. Among published studies, Galux has reported results covering the largest number of targets with top-tier performance, validating the strength of its technology.
At the center of this capability is Galux’s Drug Design Center. We sat down with the co-founders and co-heads, Jonghun and Jinsol, who lead development of the company’s core technologies, to talk about the essence of AI protein design, Galux’s vision for the future, and the people making it real.
Galux's co-founder and co-heads of Drug Design Center, Jinsol and Jonghun
The Road to Galux: Bridging Research and Industry
Q. How did each of you get into your field of study?
Jonghun: I majored in chemistry and statistics at Seoul National University and earned my PhD in physical chemistry in Professor Seok’s lab (now Galux’s CEO). My work has focused on predicting protein structures and properties and on design methodologies. I was fascinated by rules and logic—hence statistics—but I was even more drawn to pursuing truth through the natural laws that shape protein structures without human intervention.
Jinsol: I also studied chemistry at Seoul National University and completed my MS/PhD in physical chemistry in the same lab. I initially worked on predicting protein–small-molecule interactions, then gradually shifted to antibody design. During a lab tour before grad school, I tried coding for the first time and immediately felt, “This is my thing.” I became curious about how protein research and coding could come together to solve real problems, and drug discovery turned out to be the perfect intersection.
Galux's co-founders: Jonghun, Claire(CEO), Jinsol, Taeyong(EVP)
Q. What are your current roles at Galux?
Jinsol: As co-head of the Drug Design Center, I set both the short- and long-term direction for GaluxDesign, our AI-powered protein design platform, and lead the team so R&D aligns with that roadmap. I develop the software GaluxDesign requires and bridge planning and execution, figuring out how to turn the technologies we build into concrete design candidates. I’m also directly involved in AI-driven drug design projects to ensure the platform translates into real-world outcomes.
Jonghun: I’m deeply involved in developing the AI algorithms that power GaluxDesign, and I work across teams to define and drive their specific technical directions. I also lead projects in close collaboration with our Drug Development Center to evaluate and substantiate the platform’s performance.
Q. What led you to co-found Galux?
Jonghun: Our lab’s methods had been recognized globally, placing at the top in competitions like CASP and CAPRI, so I wanted to see the impact we could have in industry. In Korea, I felt our team could contribute most meaningfully to AI-enabled drug discovery and development. I also believed in the team and in our CEO’s leadership, and the chance to keep doing research in an industrial setting was decisive. People around us worried when we chose to start a company, but I had a quiet confidence that it would work out.
Jinsol: With AlphaFold pushing prediction accuracy to new heights, it felt inevitable that the era of AI-based protein design would arrive. I wanted to be at the center of that shift. Building on Professor Seok’s 20-plus years of expertise, we were steadily advancing the science, and I was curious how far an AI platform could take drug discovery and development. Publishing our work and collaborating with industry would accelerate progress through real feedback. Another big draw was continuing to work with people from our lab. I trusted the maturity of our technology and saw a real chance to win in the market.
Early days at Galux office
At the Core of Galux’s Technology: The Drug Design Center
Q. What does the Drug Design Center look like, and what strengths do team members bring?
Jonghun: Think of the center as Galux’s engine. We bring together researchers developing structure-based and physical-chemistry-driven design methodologies with those approaching problems through AI and bioinformatics. We maximize individual strengths and collaborate tightly.
Jinsol: Many team members didn’t start with prior experience in protein or drug design, but they’ve ramped quickly in both AI and biology. They contribute through intuition and creativity, and the variety of backgrounds actually creates a strong collaborative synergy.
Galux's Drug Design Center
Q. How is Galux’s AI design technology differentiated from competitors?
Jonghun: We precisely define what an AI for protein design must achieve, and our know-how permeates everything—loss functions, data processing, you name it. This comes from over two decades of physical-chemistry-based protein research in our lab. We train AI to elevate design performance by grounding it in the physical-chemical principles of protein structures. Crucially, we operate our own wet lab, enabling fast, accurate feedback on designed molecules.
Jinsol: Our philosophy also differs: rather than “bolting AI onto” the traditional process, we’re rebuilding the development protocol around AI. Because our company’s vision is AI-driven drug discovery and development, we work in a proactive environment where everyone shares the same goal and is focused on advancing the technology.
De Novo Antibody Design and Expansion to More Ambitious Targets
Q. We have to talk about the recent results. Galux attracted global attention for designing novel antibodies across multiple targets from scratch. How did you achieve this?
Jonghun: In March, we released de novo antibody design results on bioRxiv. Using GaluxDesign, we created antibodies with strong binding affinity across six targets—the highest number reported among publicly available studies to date—and that’s driven intense interest from global pharmas, partners, and investors. As our research has continued, we’ve now expanded to nine targets, and we plan to disclose those results soon.
This wasn’t something design alone could deliver; it was a company-wide effort. Our Corporate Development Division defines the target proteins, epitopes, and intended roles based on research and business strategy. The Drug Design Center then uses AI to design against those specifications. The Drug Development Center validates performance experimentally and selects the most promising antibodies. With a shared goal, each group operates at a very high level, and that’s how we achieved these results.
Jinsol: For me, it’s the most memorable project since founding the company. From day one, we envisioned designing novel antibodies with AI, and of course we failed many times at first. Instead of getting discouraged, we set out a mid-to-long-term plan, staged our goals, tuned the difficulty, and kept delivering over six months, a year, and beyond. That discipline led to success. It also reflects five years of constant communication across the company about strategy and direction.
[Full study is available here: Precise, Specific, and Sensitive De Novo Antibody Design Across Multiple Cases]
Q. What are you most focused on now?
Jonghun: After the major milestone of de novo antibody design, our goals are to improve success rates and to tackle targets that were previously out of reach for conventional experimental design. Because GaluxDesign isn’t limited to antibodies, we’re expanding into general proteins as well. We’re pursuing research that goes beyond technical demonstrations to concrete outcomes: market traction, deals with major pharmas, and the generation of bona fide drug candidates.
Jinsol: We’re drawn to problems that are unknown or hard to access. As our track record grows, we’re moving into increasingly challenging areas such as GPCRs, ion channels, and more. Real achievement isn’t just iterative redesign, but it’s succeeding in domains where no one has before. Within the next five years, it will be natural for AI to lead across drug discovery. Big Tech companies like Google are investing heavily and visible results are emerging. Galux has already secured a major milestone with de novo antibody design, and we plan to build on that foundation.
Target and modality agnostic characteristic of GaluxDesign
Ownership and Creativity, Growth through Autonomy
Q. What is the culture of the Drug Design Center like?
Jonghun: We’re proactive and flexible. Teams form around projects, and people rotate between leading and contributing roles. We value creativity and autonomy; rather than enforcing tasks, we set direction so individuals can solve problems actively.
Jinsol: There’s no heavy-handed control. People work autonomously, yet most hold themselves to very high standards. We try to create a culture where individuals leverage their strengths to achieve shared goals, and so far the synergy has been excellent.
Q. What makes your colleagues special?
Jinsol: Everyone is deeply specialized, and teams are constructed around those strengths. In AI, a single outstanding contributor can move the needle a lot, and we have more than ten such people. Even though the company isn’t widely known, many joined after digging into the technology on their own. That shared passion for the field sets us apart.
Jonghun: Our team includes people from diverse backgrounds, but they adapt quickly and contribute with strong motivation. That’s because our hiring emphasizes logical and critical thinking, deep research understanding, active debate, and self-direction. We select for high-caliber talent.
Q. What’s the most important quality for someone joining Galux?
Jonghun: If I had to choose one: a genuine desire to contribute to AI drug discovery and development. That applies across the company, not just our team—technology development, experimentation, commercialization, communications, finance and fundraising—there are many ways to contribute. In our team specifically, there are two pillars: building the AI platform and designing candidates. With AI experience, you can contribute even without domain expertise; with strong structural-biology intuition, you can learn the AI side here. The will to contribute matters most.
Jinsol: Company-wide, different roles demand different strengths, so it’s hard to name just one quality. But Galux is operating at a world-class technical level and pioneering new paths in drug discovery. We’re looking for people who don’t simply follow trends but discover new fields and enjoy creative challenges.
In Closing
Q. Finally, what do you hope to achieve?
Jonghun: I want our technology to become medicines the world truly needs, and to witness those medicines saving lives. More broadly, I hope we help cultivate a new growth industry for Korea, built on best-in-class performance. Times are tough economically and socially; I’d like AI drug discovery to open a path forward, and for Galux to lead it.
Jinsol: I feel similarly. AI will redefine drug discovery, and I want Galux to create and lead that moment. Within the company, I hope we build a place where everyone grows together and stays for the long term.
Even now, Galux’s Drug Design Center is turning next-generation design technologies into reality—pushing beyond established limits. What began as a small possibility in a university lab has grown into technology that competes on the world stage, and the journey continues.
With a philosophy that sees AI not as a tool but as the new standard in drug discovery, and with a team that collaborates naturally through autonomy and creativity, Galux’s Drug Design Center will keep turning difficult challenges into real-world outcomes, one by one.
We are shaping the future of drug discovery and development with AI, and we welcome the pioneers to join us.
[Check our open roles]
Powered by Froala Editor