Builds the Axis AI platform, a transformer-based generative and predictive model that designs regulatory DNA elements (enhancers, promoters) to enable safer, more precise cell and gene therapies with programmable control over gene expression.

Origin has publicly released the Axis model with free API access for researchers, announced plans to expand into multi-modal biological modeling (RNA, protein, cellular representations), and published benchmarks showing Axis outperforms Google DeepMind's AlphaGenome by 6.7% in regulatory element activity prediction. They've detailed a vision for a comprehensive library of validated, cell-type-specific regulatory elements and positioned Axis as a foundational platform for programmable biology beyond gene therapy.

GitHub and preprint activity suggest rapid iteration on transformer architectures for DNA sequence generation and function prediction. Hiring patterns indicate investment in wet-lab validation capabilities to build proprietary synthetic regulatory element datasets, a classic data-flywheel strategy. Conference appearances and advisor networks (MIT Broad Institute, UCSF gene therapy labs) hint at undisclosed pharma collaborations. The free API strategy mirrors early-stage platform plays designed to capture market share and generate training data before monetizing through enterprise licensing or co-development deals.