AI & predictive modelling · Falls Church, VA
We build predictive and explainable AI for government agencies and large enterprises — which storm takes the grid down, which asset a hurricane reaches first, who is about to resign, which patient is quietly getting worse. And we publish working tools, so you can check our thinking before you ever talk to us.
We have built for
§ 01
01
Prediction, record linkage and forecasting on the data you already hold — built from the start to run in production, where your team can keep it running.
02
Courses we teach live, to real rooms — analysts and the people they report to, together, with no programming background assumed of anyone.
03
A model only earns its keep when it reaches the person who has to act on it. We build the platform that puts it in front of them, in time to matter.
§ 02
Energy · Utilities
2 years
A two-stage model that forecasts damage by district, then converts that into the crews and materials to pre-stage.
A major US investor-owned electric utility
15%
reduction in outage times
Energy · Environment · Public sector
38 years of records
Roughly 62,000 incidents across 38 years. 94% were reported the same day. Only 20% were resolved the same day — a gap that had never been quantified.
Government of Alberta
~62,000
incidents analysed
Automotive · Telematics
19,000 km over 11 days
Sensor telemetry from an 11-day, 19,000-kilometre expedition, reduced to a behavioural signature that identifies who is behind the wheel.
A global automotive manufacturer
78%
identification accuracy
§ 03 — products
Most consultancies ask you to take their word for it. We would rather just show you. 22 applications are live right now — no signup, no demo call, no form. Click one and it opens. 7 of them are federal and defense tools, and we built every one from public records: federal spending data, FAA sighting reports, GAO and CRS papers, ship tracking. Nothing in them is privileged and none of it came from a client. They exist because the questions they answer are genuinely hard, and being able to answer them is what we're actually selling.
Federal & defense intelligenceNetwork · map · time slider
An interactive map of the US government’s Counter-UAS contracting network, drawn live from public federal spending records. Every node a real agency or firm, every line a real award. Built to answer a question the market rarely visualises: where do small and disadvantaged firms actually sit in Counter-UAS procurement?
Open Constellation
Applied productsSHAP · NHANES reference · batch + API
Estimates diabetes, pre-diabetes, cardiovascular, hypertension, kidney and fatty-liver risk from measurements a routine physical already produces — then shows a SHAP attribution for every prediction, so the output is a set of levers rather than a number. Each input is placed against the NHANES population distribution.
Open GlucoGuard
Federal & defense intelligence56 programs
Major defense acquisition programs, and the distance between what they were promised to cost and what they cost now. 56 verified programs, $1.3T baselined against $1.7T estimated, from public GAO, CRS and DoD sources.
Open Drift
Applied productsGeo-held-out CV · 49 states
Scores every US neighbourhood against a brand’s existing store footprint and ranks the places that look like its best-performing sites but hold no store yet. Cross-validated by holding out whole states, so the model is judged on geography it has never seen. Built on public demographics, business mix and POI density.
Open Twinsite
Science, explained56 visualisations
56 interactive visualisations across probability, statistics, calculus, linear algebra, physics, engineering, computer science, chemistry and biology.
Open SciVizHub
Applied productsLive demo, no signup
Attrition risk with the explanation attached. Open it with a one-click demo persona — CHRO, VP or manager — and see the whole thing without an account.
Open FlightRisk§ 04 — Founder
Aous is a physicist. He spent a decade in federally funded research — the Milagro observatory at Los Alamos, then a National Academy of Sciences fellowship on NASA’s Fermi mission, held across NASA and the Naval Research Laboratory. It was the same job every time: find a faint real signal buried in noise, and then defend it to a room of people whose job is to find the hole in it. That turns out to be this job too.
Every figure above is checkable in a public database. That's the point.
“Working with Dr. Abdo was a real pleasure. His abilities to extract value from big data are first class. Creative thinking, attention to detail, and high personal standards ensured that we were successful in meeting our objectives.”
Tell us the decision you keep having to make with less certainty than you would like. We will tell you honestly whether the data you already have can support it — and if it cannot, we will say so on the first call rather than the third.
