Interview with China Economic Weekly: To succeed in biomanufacturing, first learn how to "converse" with AI.

“If younger students ask me for advice on their studies, I would say to study Chinese well first,” Tan Yang, head of the Agent project at Tianwu Technology, told our reporter.

This Gen Z tech worker's answer was somewhat unexpected.

Many people perceive protein research as an extremely high-barrier-to-entry field: complex models, lengthy experiments, and expensive equipment, accessible only to large research institutions and leading companies. However, Tan Yang believes that with the rapid development of AI + biomanufacturing, this industry landscape is being rewritten.

Tan Yang and a group of young people around him are trying to make AI a “partner” in scientific research, lowering the barriers to entry.

Recently, Tianwu Technology, where Tan Yang works, released MatwingsVenus™ (晓鹜™), a conversational protein research intelligent agent. Through this agent, researchers can submit their needs as if in a conversation, and the agent will complete industry research, database retrieval, protein design, and connect automated experimental verification and result iteration. Processes that previously required multiple teams working in relay have been reconnected.

In his view, the significance of such tools is not only to improve efficiency, but more importantly, to give more young people the opportunity to get closer to the field of biomanufacturing.

Tan Yang, Head of R&D for Tianwu Technology's Agent Project

01

“R&D capabilities are becoming shareable.” Biomanufacturing is considered one of the key directions for future industries. Whether it's innovative drugs, functional foods, agriculture, or bio-based materials, the research and design of biological components such as proteins, enzymes, and microorganisms are indispensable.

Tan Yang's industry is one of the most cutting-edge directions in biomanufacturing—protein design. In the past, this was a field heavily reliant on "craftsmanship," requiring over 10 years of training for a seasoned researcher to independently design functional proteins.

But now, AI is changing the face of this industry. Through conversational AI agents, even non-professionals can complete tasks ranging from industry research and data analysis to experimental design. The AI ​​agent automatically breaks down tasks, calls up tools, and even connects to the experimental verification process.

“One important change brought about by AI is that some previously highly scarce capabilities are beginning to be accessed in a more inclusive way,” said Tan Yang. Previously, understanding a cutting-edge field might require flipping through dozens of papers and hoping for accurate comprehension. Now, simply give a paper to AI, and it can interpret it clearly in minutes.

Knowledge barriers haven't completely disappeared, but the threshold for acquiring and understanding knowledge is decreasing. At a further level, some tasks that were previously only achievable by large institutions are now becoming possible through "sharing."

A key idea behind Matwi-ngsVenus™, developed by Tan Yang's team, is to streamline the design and validation process. The platform not only helps users complete front-end research and protein design but also connects the results to automated wet laboratory workflows, allowing designs to quickly receive experimental feedback and move to the next round of optimization. This way, research and development no longer remains merely "paperwork" but can form a closed loop more quickly.

In a de novo design project targeting an immune regulatory receptor, Tianwu Technology successfully obtained dozens of novel binder molecules with in vitro cell blocking activity using this platform, completing a closed loop from design to validation.

In the complex site-specific mutation modification of the sweet protein Monellin, the platform employed a strategy of "Agent design—automated experiment—AI feedback—Agent redesign" to progressively narrow the search space and optimize the candidate set. Through continuous iterative optimization, 24 representative candidate sets were formed, resulting in several superior candidate variants. The sweetness of multiple samples was increased by more than ten times compared to the wild type, and the heat resistance remained at a high level of approximately 75°C.

These two cases are interconnected, demonstrating the practical capabilities of intelligent agents in accelerating the development of innovative protein drugs.

In the past, these tasks often required multiple teams to collaborate and switch between different tools to complete. Now, they can be reorganized on a single platform. For young researchers and entrepreneurial teams, this is not just about increased efficiency, but more importantly, it makes R&D capabilities less out of reach.

For Tan Yang, this change is an industry trend and a key reason for his choice of this path.

“My mentor, Professor Hong Liang, often says that choice is more important than effort,” he says. “Our generation is more accustomed to using tools to solve problems and believes more that technology can lower barriers to entry.”

02 Entering the Future Industry: First Learn to “Express”

Tan Yang’s mention of “learning language first” is not a joke.

He frankly states that many students with technical backgrounds, whether in biology, computer science, or other development directions, tend to focus entirely on specific technologies and localized needs, neglecting expression, communication, and the ability to translate ideas into words and tasks.

“Most of my team members have a biology background; I’m in the minority,” he says. Because of this, he has a more direct understanding of interdisciplinary collaboration and the value of “translating” requirements into actionable tasks.

In his view, as artificial intelligence gradually becomes an important tool for scientific research and industrial innovation, this ability becomes even more crucial. Because human-AI collaboration is essentially a matter of high-quality expression. "If you can't even speak clearly to AI, it certainly won't understand what you're trying to express, and it can't help you achieve what you want to do," Tan Yang said.

Further than expression, there's imagination.

"How do you paint a bigger picture for AI and then do bigger things?" Tan Yang said. With the lowering of technological barriers, what truly differentiates us will be our ability to ask bigger questions, describe bigger goals, and envision richer application scenarios. In the past, some experience might have been an advantage, but in a phase of rapid iteration of new technologies, experience can sometimes become a constraint.

"The more experience you have, the more constraints you face," he said. The advantage of young people lies precisely in their lack of path dependence and their willingness to imagine new solutions.

The reason future industries attract young people is precisely because they are still growing. Today, as AI begins to penetrate deeper into laboratories, the biomanufacturing track is also being reshaped: some are improving algorithms, some are streamlining experimental processes, and others are trying to lower the barriers to innovation, allowing more people to participate.