SK-backed startup launches Korea’s first pharma AI platform

SOUTH KOREA – A Silicon Valley-based startup, PhynX Lab, supported by South Korea’s SK Group, has launched Cheiron, a generative AI platform specifically designed for the pharmaceutical industry.

Positioned as a cutting-edge solution, Cheiron is tailored for researchers and pharmacists, offering an advanced approach to information retrieval and analysis.

The platform, described as the flagship AI solution of PhynX Lab, represents a significant step forward in leveraging AI for the pharmaceutical sector.

How Cheiron works

Cheiron’s key features include a one-stop, semantic-based search capability powered by a modular Retrieval-Augmented Generation (RAG) framework.

Unlike traditional RAG systems, the modular RAG framework allows the platform to retrieve relevant information from diverse databases, analyze intent using natural language processing, and self-verify the accuracy of its responses before delivering high-quality outputs.

Explaining this functionality, the company likened the framework to Lego blocks, noting that it can operate independently or connect logically based on specific requirements.

This structure not only facilitates adaptability to new technologies or data but also enhances security through optimized data storage selection.

The platform integrates multiple data sources, including academic platforms like PubMed, general search engines, open-source materials, and internal corporate documents, all accessible through a single interface.

Beyond data retrieval, Cheiron can automate workflows such as organizing information and generating reports.

Additionally, it supports multiple major languages alongside Korean, broadening its usability.

Addressing industry challenges

PhynX Lab highlighted Cheiron’s potential to address the limitations of existing large language models (LLMs) in healthcare.

The company explained that traditional models are constrained by their reliance on pre-learned data, often leading to misinterpretation or the generation of unreliable answers when encountering unfamiliar information.

To overcome this, Cheiron incorporates features that significantly enhance the reliability and efficiency of information processing.

During its demonstration, PhynX Lab cited findings from user tests, revealing that Cheiron can reduce document search time by up to 80% during the research planning phase compared to traditional methods.

Beyond improving efficiency, the platform also supports critical functions such as research planning, candidate substance discovery, clinical trial design, production process development, and market analysis.

PhynX Lab is now working on further enhancing Cheiron to meet the evolving needs of pharmaceutical companies.

Industry trends

Cheiron’s launch aligns with a growing trend of generative AI solutions targeted at the pharmaceutical industry.

Earlier this year, Daewoong Pharmaceutical, a publicly-listed Korean company, introduced its genAI-powered drug discovery platform, which operates on a database of 800 million compounds.

Global tech giants have also made strides in this area; Google developed the Tx-LLM by fine-tuning its PaLM-2 model for drug discovery.

IBM also partnered with Boehringer Ingelheim in November 2023 to use generative AI for identifying antibodies capable of neutralizing disease cells.