Accelerating Life Sciences Advancements
Accelerating Life Sciences Advancements
Salt enables life sciences organizations to adopt and harness AI faster than before
Salt enables life sciences organizations to adopt and harness AI faster than before
Reliable & Reproducible AI Processes
Reliable & Reproducible AI Workflows
Engineered to accommodate ongoing advancements in the scalability and flexibility of computational techniques for drug discovery - Hot swap best in class models trained for biology
A Step-Change in Speed & Efficiency
Each new model onboarded is optimized for speed when hosted on the Salt AI platform, enabling faster time to output at reduced costs of compute on comparable hardware platforms
A Step-Change in Speed & Efficiency
Each new model onboarded is optimized for speed when hosted on the Salt AI platform, enabling faster time to output at reduced costs of compute on comparable hardware platforms
Transparency within the Workflows, Collaboration at Each Step
Once the territory of computational biologists and engineers, Salt AI's visual transparency and node structure invites collaborative interaction beyond those who code. Biochemists, clinical development teams, executives, and other key stakeholders can engage real-time to shape workflows and iterate solutions, QCing output at each node.
“Since we met the team almost a year ago, Salt AI’s platform has helped us to innovate at rapid speed in designing new therapeutics for cancer.”
Dr. David Agus
CEO of the Ellison Medical Institute
“Since we met the team almost a year ago, Salt AI’s platform has helped us to innovate at rapid speed in designing new therapeutics for cancer.”
Dr. David Agus
CEO of the Ellison Medical Institute
“Since we met the team almost a year ago, Salt AI’s platform has helped us to innovate at rapid speed in designing new therapeutics for cancer.”
Dr. David Agus
CEO of the Ellison Medical Institute
Our AI Models
Beyond LLMs, we support the best-in-class life sciences research models in our always updated model library
Folding
AlphaFold2 (22x Faster)
Salt accelerates AlphaFold2 inference by 22x, significantly improving its speed. AlphaFold2 is a deep learning model that predicts protein structures from amino acid sequences with high accuracy.
Diffusion
Chroma
Chroma is an open-source, embedding database designed to make it easy to build AI applications with embeddings. It allows you to store, query, and manage embeddings alongside their metadata.
Folding
ColabFold
ColabFold simplifies protein structure prediction by providing a user-friendly interface to run complex algorithms like AlphaFold2 and RoseTTAFold in Google Colab.
Docking
Haddock
HADDOCK (High Ambiguity Driven protein-protein DOCKing) is a flexible docking approach used for modeling biomolecular complexes.
LLM
Llama3
Llama 3, also developed by Meta, is a next generation open source LLM, with improved reasoning, and code generation.
Protein Sequence
ProteinMPNN
ProteinMPNN is a deep learning method for protein sequence design, finding sequences that match a given protein backbone structure.
Our AI Models
Beyond LLMs, we support the best-in-class life sciences research models in our always updated model library
Folding
AlphaFold2 (22x Faster)
Salt accelerates AlphaFold2 inference by 22x, significantly improving its speed. AlphaFold2 is a deep learning model that predicts protein structures from amino acid sequences with high accuracy.
Diffusion
Chroma
Chroma is an open-source, embedding database designed to make it easy to build AI applications with embeddings. It allows you to store, query, and manage embeddings alongside their metadata.
Folding
ColabFold
ColabFold simplifies protein structure prediction by providing a user-friendly interface to run complex algorithms like AlphaFold2 and RoseTTAFold in Google Colab.
Docking
Haddock
HADDOCK (High Ambiguity Driven protein-protein DOCKing) is a flexible docking approach used for modeling biomolecular complexes.
LLM
Llama3
Llama 3, also developed by Meta, is a next generation open source LLM, with improved reasoning, and code generation.
Protein Sequence
ProteinMPNN
ProteinMPNN is a deep learning method for protein sequence design, finding sequences that match a given protein backbone structure.
Our AI Models
Beyond LLMs, we support the best-in-class life sciences research models in our always updated model library.
Folding
AlphaFold2 (22x Faster)
Salt accelerates AlphaFold2 inference by 22x, significantly improving its speed. AlphaFold2 is a deep learning model that predicts protein structures from amino acid sequences with high accuracy.
Diffusion
Chroma
Chroma is an open-source, embedding database designed to make it easy to build AI applications with embeddings. It allows you to store, query, and manage embeddings alongside their metadata.
Folding
ColabFold
ColabFold simplifies protein structure prediction by providing a user-friendly interface to run complex algorithms like AlphaFold2 and RoseTTAFold in Google Colab.
Docking
Haddock
HADDOCK (High Ambiguity Driven protein-protein DOCKing) is a flexible docking approach used for modeling biomolecular complexes.
LLM
Llama3
Llama 3, also developed by Meta, is a next generation open source LLM, with improved reasoning, and code generation.
Protein Sequence
ProteinMPNN
ProteinMPNN is a deep learning method for protein sequence design, finding sequences that match a given protein backbone structure.
Biology. Sooner.
Advance your understanding of biological processes faster with flexible and performant AI.
Biology. Sooner.
Advance your understanding of biological processes faster with flexible and performant AI.