Bordeciel
We build decision-grade software for the teams operating in complex private markets.
Manifesto
Every market throws off more signal than the people inside it can hold. Bordeciel builds systems that ingest that signal, resolve it into structure, and return decisions the client can act on. What we ship changes by sector. The operation underneath does not.
Underneath, the method is mathematical. Statistics, information theory, and applied inference, tools that measure their own uncertainty. Every model runs against a baseline. Every claim carries an error bar. That is what "engineered" means when we say it.
Projects
Three product lines. Each enabling a client to automate a role and to eliminate uncertainty in data they rely on for day-to-day operations. Each is a place the engine has been productized against a real buyer in a real vertical. Every deployment sharpens the base layer the next one runs on.
Engine, market intelligence automation
The Engine reads every public trace of a market, directories, registries, filings, maps, digital footprint, social, and resolves them into a ranked universe of prospects. Point it at a sector it has never seen and it re-composes itself around the shape of that sector's public and private data. Where an analyst samples the market once a quarter, the Engine re-runs the market on demand. Every sector it runs against widens what the engine's base layer knows about the private economy.
Valuence, content valuation modeling
Valuence turns creator performance into per-post dollar value. Live engagement, audience quality, and cadence signal from Instagram, YouTube, and TikTok run through a statistical model and come out as a number both sides of a negotiation can point at. Creators use it to know their worth. Agencies use it to close. The creator economy, instrumented the same way the rest of the engine instruments its verticals.
Systems, bespoke operational platforms
Systems is where the engine meets the field. For a Fortune 500 client we built the operating environment its field team actually opens: a mobile-first surface where prospect ranking, territory potential, and market segmentation from the engine reach the operator on the road, and where every field interaction feeds back as ground-truth. Some verticals still lack software that matches how they operate; we build it, and it becomes both the interface layer for the engine's outputs and the first-party sensor that keeps the engine honest.
Methodology
We build systems the way experiments are run, and we build the engine the way instruments get more accurate: one deployment at a time, against a shared base layer.
how we build systems
Every model has a baseline. Every deployment has a control. Every version is measured against the last. Every claim ships with its confidence score. When the environment changes, the system adapts and delivers consistency.
how the engine compounds
Every vertical deployment feeds the base layer. Signal harvested for one client resolves into structure the next client's model runs against. Field ground-truth from an operating team calibrates what the scrapers see the following week. The engine gets sharper every time it is used, in a way any single product line, on its own, could not.
who reads the engine
Public markets are measured tick by tick. The private economy, the businesses that make up the majority of employment and output, is not. What the engine's base layer accumulates, across every vertical it is deployed in, will eventually be read by the institutions that need it most. Financial institutions ingesting it as signal for models built on the private sector. Public institutions writing policy in tune with the businesses it affects. We are not there yet. The path there is one vertical at a time.
Founders
Steven, Infrastructure
Ran quant infrastructure at RBC and, before that, Morgan Stanley, the kind sophisticated models need to survive production under load. The difference between a research artifact and a real system.
Celestino, Realtime
Engineered real-time systems at Microsoft. Reached Microsoft AI via the Nuance Labs acquisition, on the Innovation Team building healthcare AI under HIPAA. Where the constraints are hardest, regulated data, real-time streams, life-critical decisions, the design has to be cleanest.
Baptiste, Innovation
Built the system that runs L'Oréal's field operations in Canada, from architecture to AI assistant, live in production for years. Finance and mathematics background, applied to creative product decisions.
Clients
- multinationals, category leaders instrumenting their market
- agencies, defending a number in every room
- financial institutions, private-sector signal, into the model
- consulting firms, analysis they cannot produce by hand
- government institutions, evidence for the public record
- private capital, opportunity mapping at fund scale
Contact
tell us the shape of the problem. we respond within two business days.