We operate at the upstream layer of clinical development, where the most consequential toxicity failures form long before they appear in datasets, animal models, or early clinical signals.
Our core technology is the A‑Flux Engine — a new computational class for toxicity insight in Pre‑IND, IND, and early‑phase clinical trials. A‑Flux computes the failure‑defining conditions that determine whether a therapy advances safely or becomes a catastrophic risk.
Traditional probabilistic and deterministic models cannot resolve these conditions. They lose resolution in the biological gaps where data is limited, complexity is high, and early‑phase evidence is sparse. A‑Flux does not infer or approximate. It computes.
Powered by advanced machine learning bounded by first‑principles reasoning, A‑Flux surfaces upstream toxicity conditions that existing paradigms cannot detect — the conditions that shape the fate of a therapy before decisions lock in.
What Ifs Tech operates at this upstream decision layer: providing high‑resolution, in silico toxicity evidence for Pre‑IND packages, IND submissions, and early‑phase go/no‑go decisions. This clarity protects programs, patients, and portfolios from downstream failures that would otherwise remain invisible until it’s too late.
A world where the life sciences don’t just simulate failure, but search deeper than today’s probabilistic and deterministic tools can reach — resolving toxicity risks upstream, inside computation, long before they can take form in humans.
What Ifs Tech’s A‑Flux Engine makes this possible by computing pre‑event biological futures: the failure‑defining toxicity conditions that cannot be measured, observed, or inferred through existing models. These conditions live in the gaps where biology is complex, data is limited, and traditional approaches lose resolution.
We push beyond those limits because safety demands more than surface‑level prediction. We believe the most important failures are the ones you catch before they exist — the ones you prevent rather than discover.
This is the transition: from reconstructing what went wrong to computing what will go wrong early enough to change it.
We are building the safety infrastructure that allows high‑stakes therapeutics to move forward with clarity, confidence, and human‑relevant certainty.
Founder & CEO | PhD Candidate | Regulatory Sciences Strategist
Esther Cashbaugh is the Founder and CEO of What Ifs Tech, a simulation‑engine company born from a personal N‑of‑1 that exposed a truth the industry wasn’t built to handle: the most dangerous toxicity failures don’t appear in data, trials, or post‑market surveillance. They form upstream, silently, long before biology has the chance to reveal them.
While pursuing her PhD, Esther encountered a pattern that would define her life’s work — a gap between how fast science moves and how slowly safety reveals itself. It wasn’t a statistical problem or a regulatory problem. It was a structural one: biology fails in ways that cannot be observed, inferred, or predicted using the probabilistic and deterministic tools the industry relies on.
That realization became the foundation for What Ifs Tech.
Drawing on her background in regulatory sciences, systems architecture, and evidence‑driven decision logic, Esther recognized that the only way to close this gap was to compute the futures that biology cannot show us. Not by extrapolating from the past, but by simulating the failure‑defining toxicity conditions that shape the fate of therapies before they ever reach humans.
This insight led to the creation of A‑Flux — a new computational class designed to compute upstream toxicity formation in the places where evidence is scarce and traditional models break.
Today, Esther leads What Ifs Tech with the same clarity that came from her N‑of‑1: safety must be computed before it exists, not reconstructed after it fails.
Her work is reshaping how the life sciences understand risk — turning the unseen into something computable, and giving clinical development teams access to biological futures that were previously unknowable until it was too late.