Healthcare

Scaling cancer therapies with AI-powered design

Key Insights 

Built an AI-powered platform to accelerate the design of RNA oncolytic viruses, reducing research friction and enabling faster iteration in cancer therapy development.

About the Client

Guska is a biotech startup focused on developing synthetic RNA oncolytic viruses that selectively target and destroy cancer cells while activating a strong immune response. Their work sits at the intersection of synthetic biology, genomics, and immuno-oncology.

Read the press release here

The Challenge

Designing oncolytic viruses is a highly complex and time-consuming process. Researchers face long iteration cycles, fragmented data across public and proprietary sources, and limited tooling to efficiently generate and evaluate new viral RNA sequences.

Guska needed a unified, AI-driven platform to centralize data, accelerate scientific discovery, and support the generation and evaluation of next-generation viral candidates, without compromising biosafety or scientific rigor.

Marvik’s Approach

We designed and built an end-to-end AI platform tailored to Guska’s research workflow, combining data infrastructure, advanced search, and generative modeling. Key components included:

  • Centralized data platform integrating public and proprietary datasets, providing structured access to genomic sequences, proteins, experimental results, and metadata.
  • Semantic search engine powered by state-of-the-art language models, enabling researchers to retrieve relevant scientific papers and evidence using natural language queries.
  • AI-driven virus candidate generation prototypes, combining generative models with scoring mechanisms and biosafety validation rules to support early-stage design decisions.
  • Modular and extensible architecture, designed to evolve alongside ongoing lab experimentation and future therapeutic programs.

The Results & Impact

  • Significantly faster design cycles for RNA oncolytic virus candidates.
  • Streamlined research workflows, reducing manual data exploration and context switching.
  • Immediate access to relevant scientific evidence through natural language search.
  • A strong technical foundation for next-generation cancer therapies, ready to scale as research progresses.

With this platform in place, Guska’s researchers can focus more time on scientific insight and experimentation, and less on data fragmentation and manual discovery.

Why This Matters

By combining AI, generative modeling, and robust data engineering, Guska is accelerating the path from scientific hypothesis to viable cancer therapies. This project shows how production-grade AI can move beyond experimentation and become a core engine for innovation in biotechnology and life sciences.

Every AI journey starts with a conversation

Let's Talk
Let's Talk