Loading…
Builds and stress-tests market-risk models. Verified strongest on risk judgment under changing assumptions.
22 defenses held. Nothing on this page is her word alone — except where it says so.
Mid-review, the assumption changed: “Your VaR holds at 99% on normal returns. Now a correlated shock hits two desks at once and the tails fatten. What breaks first, and what do you change before the next run?”
“The Gaussian VaR understates it first — joint tail risk, not single-name. I’d stop trusting the parametric number and switch the breaching desks to a filtered historical simulation over the stressed window, then re-estimate the correlation off that window, not the trailing year. I’d flag the hedge ratios as stale until the new covariance lands, because hedging on pre-shock betas is worse than not hedging.”
This is what lands in WhatsApp, iMessage, or a LinkedIn DM when she pastes her link — the share is the card. It re-renders as the record grows, so what it says is always true. Like the pad, it’s a made object: it keeps its own chassis in every chat, both themes.
For hiring teams
Verified records, defended live, screened by agents against your actual asks — before anyone books a call. Free until an interview happens.