Reticular Launches: Steerable and Interpretable AI for Protein Engineering

Reticular recently launched!

‍

^{"Pioneering mechanistic understanding to guide biological AI systems with limited data"}

‍

^{TL;DR: Reticular gives pharma companies precise control over protein AI models, enabling reliable drug discovery without millions in wasted experiments. They are unlocking encoded information in these models through foundational AI interpretability research, making steering as easy as prompting ChatGPT.}

‍

Founded by Nithin Parsan & John Yang

Nithin and John are huge AI + Bio nerds who:

• Met 7 years ago competing in International Biology and Neuroscience Olympiads and were roommates at MIT for 4 years
• Published ML/bio research in NeurIPS, Nature, and PLoS ONE

‍

The Challenge:

Information in biology is incredibly scarce and expensive to validate. While protein AI models like AlphaFold have revolutionized drug discovery, they're still black boxes:

• Companies waste millions testing AI-generated designs and they can't easily control what these models output
• Limited data makes validation slow and expensive

‍

Reticular's Approach:

Instead of trial-and-error, the team is leveraging mechanistic interpretability techniques that excel at extracting sparse knowledge from models even with scarce data. Reticular is demonstrating that the same advances from frontier AI labs in steering models like Claude are applicable to unlock protein language models.

For their design partners, they are delivering:

• Direct steering of protein models towards desired properties
• Interpretable biological features backing every design
• Efficient exploration of combinatorially massive design spaces with limited data

‍

See It In Action:

‍

They are steering Green Fluorescent Protein towards more fluorescent sequences by directly controlling the model's internal knowledge.

Learn More at reticular.ai, the proof-of-concept blog post, and at demo.reticular.ai.

‍

The Team Is Looking To Connect With:

• Startups working with biological language models for design partnerships
• Pharma teams building generative discovery pipelines
• Researchers working on AI interpretability

‍

Learn More

‍

^{🌐 Visit www.reticular.ai to learn more.}

^‍

^{⚡ Interested? Partner with Reticular here.}

^‍

^{🤝 Working with biological AI or have relevant connections? They would love to chat! Schedule some time here or email them here.}

^{‍
👣 Follow Reticular on LinkedIn & X.}