Skip to content
Neil Sethi

Selected work

1health (Tachin.AI) · Jan 2026 – present

The platform behind new clinical products

Director of Platform Product and Partnerships

Every new clinical product needs the same regulated infrastructure: patient data, lab ordering, eligibility and compliance. I built 1health's developer-facing products: the portal, Patient Vault (a headless API over 2.7M patient records) and MCP tooling that lets AI coding agents discover and call the APIs directly.

Visit the developer portal

Overview

1health is a HIPAA-compliant, SOC 2 Type II platform that healthcare products are built on. I own the developer-facing side: how engineers, and increasingly their AI coding agents, discover, understand and call the platform's APIs.

The problem

Every team building a clinical product hits the same wall: eligibility screening, CLIA lab ordering, medical-director oversight, results delivery and chain of custody. Building it yourself takes months to years and a team of hires, and none of it differentiates the product.

There is a second, newer problem. Much of the integration work is now done by AI coding agents. Agents are becoming primary customers of the platform, and they need a front door designed for them.

What I built

  • The developer portal at dev.1health.io.
  • Agent-readable documentation, and MCP tooling that lets a coding agent find the right endpoint and call it directly.
  • Patient Vault, a headless patient data API over the secure data layer holding 2.7M patient records.
  • Lab ordering through national lab networks for startup and enterprise build partners.

The product bet

I treated AI agents as first-class users of the platform, not an add-on. Documentation is designed to be read by a model, APIs are designed to be discovered by one, and the portal serves both audiences from the same source.

I run this site the same way.

See /llms.txt

What happened

Partners ship compliant products in weeks rather than months.

Muse Bio, a regenerative medicine company working with stem cells, ran its donor program on Google Forms, spreadsheets and paper, out of one San Francisco clinic. It launched an automated program in about a month: eligibility and consent, lab orders, physician approval, specimen tracking and results. The program is now ready to scale nationwide. Their COO said that without 1health they would probably have needed to hire 10 people.

Expert-Dx, a diagnostic consultation service for physicians with hard lab cases, led by Dr. Michael Laposata, went live in about a month. Getting from decision to a usable product normally takes about twelve.

Muse Bio case study Expert-Dx case study