Psychiatry prescribes without an objective signal for who will respond. Alto’s thesis was that such signals exist and can be measured — biomarkers drawn from EEG, wearables, and cognition that identify biomarker-positive patients and match the right drug to the right person. Joining meant, after years of building things without a target use case, working on something with a real product line in an area where treatments are genuinely inadequate.

Roughly the twenty-fifth employee, through to the IPO in February 2024 and on the floor of the New York Stock Exchange for the bell.

Five years of work across the full width of that thesis: an event-related-potential biomarker program built from nothing to a Phase 2 primary pharmacodynamic endpoint; discovery and validation pipelines over EEG, ERP, wearable, speech, and cognitive data from trial cohorts and legacy datasets; acquisition pipelines and QC dashboards used to onboard clinical sites and triage data quality during live collection; translational rodent electrophysiology; and, latterly, three production LLM tools for trial operations. First-author posters at SOBP 2025 and ACNP 2026.

What changed over those five years is where the leverage was. It started in signal processing — getting a clean measurement out of a noisy one. It ended in tooling: the constraint was rarely the analysis, it was everything wrapped around the analysis, and that is what the current independent work is pointed at.