An operating system for institutional forecasting

Build the future you'll operate in. Not the one that arrives.

Famine. Energy. Materials. The decisions ahead are forecasts in disguise. Anominal AI deploys inside your boundary as a dedicated instance — five models compete on every series, your data trains your models, your infrastructure runs them.

Book a Discovery call »

Built by the Senior AI Technical Advisor to the World Bank Group · Validated under IMF sponsorship at the central bank level · Close customer engagement with Of The Earth Group across hemp genetics, cultivation, and finished textile · Sovereign by design

What the platform does

Five models compete on every forecast. The cone of probable outcomes is the product.

A curated stable of forecasting models — classical, modern, and frontier — runs in parallel on every series. Composite scoring resolves them into a single calibrated forecast. Walk-forward backtesting against your own historical data validates accuracy before any forecast drives a decision.

Live forecast visualization · Coming soon

The methodology validated at sovereign Treasury scale, applied at operational granularity. Calibrated confidence intervals, multi-model resolution, full evidence trail.

The platform installs as a dedicated instance inside your boundary; models install, replace, and upgrade like applications. Your data, your trained models, your infrastructure — never multi-tenant, never aggregated, never shared. The depth of bespoke curation per engagement creates unparalleled models for our customers.

One platform. Three pillars. Built for institutions that need to trust their own tools.

Foresight

Global market data, fused with yours, across every horizon that matters, from the next two minutes to the next century.

We orchestrate and blend your data with curated signals from the world's markets, at your preferred intervals; tuned to the decisions you actually make. The curation is the product. Five forecasting models run in parallel on every series. A calibrated range of probable outcomes, never a single number pretending to be certain.

Integrity

Every forecast is backtested against your own historical data before it drives a single decision. The methodology is designed to be interrogated, not accepted on faith.

Five models competing on every forecast. Calibrated confidence intervals. A full evidence trail from source data to output. The methodology that satisfied institutional auditors will satisfy yours.

Sovereignty

Your data runs on infrastructure built exclusively for you. It never touches another client's environment. If you end the engagement, everything we built together stays with you.

Dedicated hardware. Open standards. Full portability. No vendor lock-in, ever. For government clients, sovereignty extends to the four dimensions recognized by international institutions: territorial, operational, technological, and legal.

The proof

96% directional accuracy on a six-month forward horizon. Central Bank of Somalia Treasury Single Account cash flows. IMF-validated, 2020 through 2023.

The same methodology is now available to sovereign clients, family offices, and enterprise operators worldwide.

Three categories of organization. One platform built for each.

Live customer engagement · April 2026 Agriculture · from seed to supply chain
Your harvest is worth more at planting than at market.

Forecasting infrastructure for growers, processors, fiber producers, and the operators running global supply chains. Of The Earth Group is in close customer engagement across hemp genetics, cultivation, and finished textile, building toward pricing forward production at or before seed planting.

See the agriculture engagement »

On Monday morning, you open intelligence built from your own data, validated against your own history, running on infrastructure that belongs exclusively to you.

You see the probable range of outcomes across every time horizon that matters, and you act before the market tells you to.

AI does not build the future. Operators with empathy, trust, and regenerative purpose do.

Bartosz and Nicole Solowiej, Co-Founders of Anominal AI

Bartosz Solowiej, Co-CEO & Co-Founder. Designed the whole-of-government PFM architecture for the Philippines Department of Budget Management. Authored the ML audit case selection methodology now in production at Armenia's State Revenue Committee. Delivered $1.5B+ in IMF and World Bank project value over seven years as Senior AI Technical Advisor to the World Bank Group. Co-author of the foundational WBG paper on AI in the Public Sector. Silicon Valley operator since 2007, contributing to two companies acquired by Flextronics and HPE.

Nicole Solowiej, Co-CEO & Co-Founder. Managed institutional audit and assurance portfolios at PwC across seven years as a chartered accountant. Scaled two operating companies past $1M in annual revenue as CFO and COO. Built the engagement methodology that makes a dedicated-instance enterprise AI platform deliverable at seed stage. Leads commercial, operations, and client engagement across every Founding Client.

Anominal AI is the first product platform of Ember Venture Group Pte. Ltd., a privately-held Singapore company founded and led by Bartosz and Nicole Solowiej, with equity held by the founders and a syndicate of active investor-advisors spanning global business development, family office management, agricultural legacy, sustainable materials, and media at the world's largest enterprises.

One conversation. Thirty minutes. Your data, not ours.

The Discovery call is not a pitch. We ask what decision you are trying to make, what data you already have, and what signals would change how you act. You leave with a data readiness assessment scoped to your operation, whether or not we work together.

Book a Discovery call »

Invest in us. Join our team.

We are open to exploring relationships with the best and brightest investors, advisors, and leaders across all disciplines. Review the opportunity »