Overview

Profluent was founded by Ali Madani, a scientist with a background in machine learning applied to biology. The company emerged from research demonstrating that the same types of large language model architectures that power natural language processing could be trained on the "language" of proteins—the sequences of amino acids that make up biological molecules. Just as language models learn grammar and meaning from vast amounts of text, protein language models learn the underlying patterns, structures, and functional rules encoded in the sequences of millions of natural proteins.

By learning these deep patterns, Profluent's models can generate entirely new protein sequences that fold correctly and carry out desired functions. This approach represents a shift away from traditional, slow, and labor-intensive methods of protein engineering, which have historically relied on trial-and-error mutation, directed evolution, and iterative laboratory screening. Profluent aims to make the design of proteins faster, more precise, and more scalable through computational generation.

What the Company Does

At its heart, Profluent builds and trains generative AI models on massive datasets of protein and genomic sequences. These models are used to design proteins with tailored properties for a variety of applications, most prominently in gene editing and therapeutics. The company's work combines several disciplines:

• Machine learning and large language modeling
• Structural and functional biology
• Genome and protein sequence analysis
• Wet-lab experimental validation

Rather than simply predicting the structure of existing proteins, Profluent focuses on the generative side of the problem: creating new proteins from scratch that meet specified design criteria. This generative capability is what distinguishes the company's approach and allows it to explore regions of "protein space" that natural evolution has never reached.

Key Achievements and Products

OpenCRISPR

One of Profluent's most significant and publicly notable achievements is the creation of AI-designed gene editors, released under the name OpenCRISPR. CRISPR systems are molecular tools used to edit DNA, and they are typically derived from naturally occurring bacterial defense mechanisms. Profluent used its generative protein models to design entirely new CRISPR-like proteins—gene editors that were not copied from nature but generated by AI trained on the diversity of known CRISPR systems.

The flagship result, often referred to as OpenCRISPR-1, is described as the first AI-generated, open-source gene editor demonstrated to successfully edit human DNA. The proteins produced through this effort differ substantially from natural editors such as SpCas9 in their amino acid sequences, yet retain functional gene-editing capability. By releasing these designs openly, Profluent sought to democratize access to gene-editing tools and encourage broader innovation in the field, positioning the work as a demonstration of what AI-driven protein design can accomplish.

ProGen and Protein Language Models

Profluent's technology builds on a lineage of protein language models, including work associated with the ProGen family of models. These models are trained to generate functional protein sequences across many families of proteins, learning the statistical and structural regularities that govern how proteins fold and function. The company has continued to advance these models, scaling them and applying them to increasingly complex design challenges, including multi-component molecular machines like CRISPR systems.

Services and Business Focus

Profluent positions itself as a company developing AI platforms for protein and biomolecule design, with applications that span:

• Gene editing and genome engineering tools
• Therapeutic protein development
• Enzyme and functional protein design
• Research collaborations and partnerships within the biotechnology and pharmaceutical sectors

The company's value proposition rests on the idea that AI can dramatically compress the timelines and costs associated with discovering and optimizing useful proteins. Where conventional approaches might require years of laboratory experimentation, generative design allows for the rapid proposal of promising candidates that can then be validated and refined experimentally.

Profluent combines its computational engine with in-house laboratory capabilities, using experimental data to continually train and improve its models. This feedback loop between AI prediction and wet-lab validation is central to its methodology, ensuring that generated proteins are not merely theoretically plausible but genuinely functional.

Significance in the Industry

Profluent is frequently cited as a leading example of the emerging field of generative biology, in which AI is used not just to understand biological systems but to actively design new ones. Its OpenCRISPR release attracted considerable attention in both the scientific and broader technology communities as a concrete demonstration that AI-generated proteins can perform real, useful biological work.

The company's approach reflects a larger trend in the life sciences, where advances in machine learning—particularly the transformer architectures underlying modern language models—are being adapted to biological data. By treating proteins and genetic sequences as a language to be learned and generated, Profluent contributes to a vision of programmable biology in which desired molecular functions can be specified and then computationally designed.

Funding and Backing

Profluent has raised venture capital funding to support its research and development efforts, drawing investment from firms interested in the convergence of artificial intelligence and biotechnology. This backing has enabled the company to build out its computational infrastructure, expand its scientific team, and pursue ambitious projects such as its AI-designed gene editors.

Summary

Profluent is a US-based company pioneering the use of generative artificial intelligence for protein and biomolecule design. By training large language models on the sequences of proteins and genetic systems, it generates novel, functional molecules—most notably AI-designed CRISPR gene editors released through its OpenCRISPR initiative. The company aims to accelerate and expand the possibilities of protein engineering, with applications ranging from gene editing to therapeutics, and it stands as a prominent representative of the growing field where machine learning and molecular biology intersect.