We're building an AI-first platform that brings decades of machine learning expertise to every stage of a data science project — from problem definition through deployment and monitoring.
Get in TouchBuilding reliable predictive models remains slow, expensive, and error-prone. Current platforms are universally recognized as inadequate.
Full development through deployment in regulated industries takes months of highly skilled specialist effort.
Current platforms bundle aggressive pricing with black-box approaches that fail model risk management requirements.
Existing market players are burdened with pre-LLM architectures. A greenfield AI-first design has a structural advantage.
An AI-first system where LLM agents handle reasoning and orchestration while rigorous classical ML engines handle prediction.
Agents guided by local and (optionally) frontier lab LLMs will embody best practices as distilled from our extensive experience. Human in the loop is always expected but can be bypassed.
GRPO and other RL post-training on challenging data science problems to learn best practices. This central open-weights RL-trained LLM enables fully on-premises execution for customers requiring complete data and technical privacy.
Gradient boosting (the GBM) outperforms neural networks on structured enterprise data. We built the first GBM and other foundational ML technology and subsequently refined these tools including CART, Random Forests, and MARS — working directly with their inventors. Part of our mission is to radically improve the performance and usefulness of legacy tools.
Unmatched depth in machine learning for structured data — direct collaboration with the inventors, 17 competition wins, 30 years of deployment.
PhD Economics, Harvard. Founded Salford Systems in 1982; led for 30+ years as a self-funded enterprise serving 300+ major corporate clients. Worked with Breiman and Friedman to commercialize CART, Random Forests, and GBM. Acquired by Minitab in 2017.
Professor of Statistics, Stanford. Inventor of Gradient Boosting, MARS, and Projection Pursuit. Co-author of the CART monograph. One of the most cited researchers in machine learning history.
PhD Statistics, Yale. Extensive enterprise analytics experience consulting to American Express, JPMorgan, with significant tenures at Capital One and Bank of America. Architect of the data science lifecycle framework guiding the agentic system.
Principal AI Systems Engineer at Virtek Vision International — four years leading production AI vision systems. Deep experience deploying AI at industrial scale.
PhD in multi-agent reinforcement learning. Postdoc at Mila (Quebec AI Institute) on LLMs for adaptive reasoning. Deep expertise in RL and large language models.
Chief Data Preparer across virtually all major Salford Systems engagements, 1990–2017. Data master for the team's 17 modeling competition victories.