Nathalie joined SlotsGamblers in March 2026 as the Data & Methodology Auditor, brought on once the desk was fully staffed and generating enough figures across licensing, games, bonuses, payments and RG that the ratings needed a formal model instead of a spreadsheet. She came to the desk from a data-analysis background, where the job was the same underneath: take numbers from several sources, check that they agree, and be honest when they don't. She's French-Canadian, based in Canada, and her job is downstream of everyone else's — Claire's RTP figures, Marc's payout timings, Kevin's bonus EV maths and Andrew's RG flags all feed into her model, and she's the one who chases a desk when their numbers don't line up before a score goes anywhere near a review.
In practice that means she doesn't wait for a finished review to check her inputs — she pulls the raw figures straight from each desk's own notes as they're logged, so a discrepancy surfaces while the reviewer still remembers the context, not weeks later when the page is already close to publishing. The first few months on the desk were mostly building the checklist-to-score pipeline itself: turning 64 separate yes/no and graded answers into one number that a reader can trust without having to read the underlying 64 lines herself. That meant deciding, category by category, what should move a score a little and what should be able to move it a lot — a licence irregularity or an active RG complaint can cap a score outright, while a slower-than-average payout on one method nudges it.
She also owns the ongoing check on her own work: she's set a quarterly cadence for auditing the model against the operator-complaint dataset the site has built since it launched in September 2025, checking that a high score still tracks a low complaint rate and adjusting the weighting when it drifts. She's honest that the dataset is still young — under a year of complaint records — so those weightings are provisional and get revisited every audit rather than treated as settled; the first quarterly audit ran in June 2026, and it confirmed the model was broadly tracking complaints correctly but flagged that payment-speed inputs were carrying slightly more weight than the complaint data actually justified, a note she's carrying into the next audit rather than changing on a single quarter's evidence. Nothing she publishes counts as final — David still gives every review the last sign-off, but he doesn't get a score to sign off until Nathalie's model has run.
The part of the job she talks about least, but that eats the most of her week, is the chasing itself: pinging Marc for a payout timing that hasn't landed yet, or asking Kevin to re-check a wagering figure that looks off against the operator's own terms page. It's rarely dramatic — most gaps close within a day once the right desk knows about them — but she's built the habit of never filling a blank cell in the checklist with an assumption, even a reasonable one, because a guessed number is exactly the kind of thing that survives quietly in a model until an audit catches it months later. When a figure genuinely can't be sourced in time, the checklist stays marked incomplete rather than the model inventing a placeholder, and the review waits.