Atomic AI

Atomic AI authors' ATOM-1 preprint reports a completed RNA secondary-structure evaluation on PDB, ArchiveII and bpRNA-1m TS0, comparing embedding probes against RNA-FM, CONTRAFold, RNAFold and a no-embedding baseline. This is a separate scientific prediction task from the previously recorded mRNA-stability evaluation; it is retrospective research, not clinical or commercial deployment.

Atomic AI reports a completed retrospective ATOM-1 evaluation of in-solution mRNA stability, outperforming 1,600 comparison methods, using a foundation model trained on in-house chemical mapping measurements.

Atomic AI is applying AI-driven RNA structural modeling to therapeutic discovery, but public large-scale commercial deployment remains early.

Atomic AI's ATOM-1 announcement describes training on in-house chemical mapping of millions of RNA sequences and more than a billion nucleotide-level measurements. This concrete experimental data and model method is distinct from a generic RNA foundation-model description.

Atomic AI develops foundation-model and structure-prediction approaches for RNA and molecular design.

Atomic AI develops foundation-model and structure-prediction approaches for RNA and molecular design.

Atomic AI's January 2023 launch release reports a $35 million Series A led by Playground Global, following a $7 million seed round.

Atomic AI identifies machine-learning leadership, ML data-generation leadership, RNA scientists, cheminformatics scientists and a Head of Business Operations, Esta Kor, alongside its founder.

Atomic AI combines computational modeling with experimental biology capabilities for iterative discovery.

Atomic AI's December 2023 author preprint describes supervised sequence-to-sequence transformer training on chemical mapping reads, followed by small probe networks on encoder embeddings and baseline comparisons. These concrete representation-learning and evaluation methods support an inference of transferable ML skills, while RNA assays remain domain-specific; no actual career mobility is demonstrated.

ML and structural modeling skills transfer broadly, while RNA and wet-lab domain expertise limits full cross-market portability.

ML and structural modeling skills transfer broadly, while RNA and wet-lab domain expertise limits full cross-market portability.