Tech Lead

Trellis AI•San Francisco, CA

9d•Onsite

About The Position

Trellis is building an automated system to streamline the complex process of getting a patient from a prescription to starting treatment. This involves automating tasks traditionally handled by intake coordinators, benefits coordinators, and prior authorization specialists. The company's technology automates interactions with electronic health records (EHRs), insurance portals, and direct desktop software, aiming for a "autopilot" mode where 90% of scripts run end-to-end without human intervention. The role of Tech Lead is crucial in scaling this system, which currently processes approximately 30 million scripts annually and $10 billion of therapies. The Tech Lead will be responsible for setting the technical direction, improving the development environment, ensuring the scalability of the infrastructure, and maintaining the reliability of the "fleet" of thousands of concurrent browser and desktop sessions that drive the automation. The company has experienced rapid growth, moving from 6 to 8 figures in revenue in less than a year, and plans to double the engineering team size by the end of the year. The Tech Lead will be a player-coach, still writing code while guiding the team. Key areas of ownership include defining the ideal development environment, managing the "fleet" of Windows sessions, implementing self-healing mechanisms for the automation, managing knowledge about payer policies and customer SOPs, and converting customer procedures into working agents. The role also involves identifying scaling limitations in the current stack (TypeScript, AWS, Temporal, Windows Server, RDS) and determining future solutions.

Requirements

Experience taking a company from zero to nine figures in revenue in two or three years.
Experience in a player-coach role, still writing code while leading a team.
Proficiency in TypeScript, AWS, and Temporal.
Experience with Windows Server and Remote Desktop Services (RDS).
Ability to diagnose runaway processes, memory leaks, and zombie sessions on live hosts.
Experience handling loopback and local RDP architectures.
Willingness to shadow sessions to debug agents mid-run.
Experience maintaining golden images.
Understanding of how to manage a fleet of thousands of concurrent browser and desktop sessions running around the clock.
Ability to define what "as close to prod as it gets" means for a complex production environment.
Experience with self-healing systems or a strong aptitude for developing them.
Ability to understand and manage complex knowledge bases for agents, including payer policies and clinical guidelines.
Experience converting unstructured information (documents, screen recordings) into functional software agents.
Ability to identify scaling limitations in a technology stack.
Experience developing robust test suites.
Experience enforcing coding practices.
Willingness to learn the rules that determine payer approval for drugs.
Willingness to work closely with operations people whose manual work is being removed.
Ability to work in the office five days a week.

Nice To Haves

Experience in healthcare data science or related fields.
Familiarity with Electronic Health Records (EHRs) and their limitations.
Experience with insurance portals and their workflows.
Background in AI or machine learning.
Experience with document automation.

Responsibilities

Set the coding bar for the engineering team.
Make high-leverage technical calls on behalf of the team, including vendor selection.
Own how engineers write tests and ship code.
Define and implement an ideal development environment that is as close to production as possible.
Diagnose and resolve issues within the fleet of thousands of concurrent browser and desktop sessions, including runaway processes, memory leaks, and zombie sessions.
Handle loopback and local RDP architectures.
Shadow sessions to debug agents in real-time.
Maintain golden images for session hosts.
Develop self-healing mechanisms for the automation system to adapt to changes in insurance portals, payer policies, and customer SOPs.
Determine what an agent should know, where that knowledge comes from, and how to prevent it from becoming stale.
Convert customer Standard Operating Procedures (SOPs) from documents and screen recordings into working agents that can handle various payers and drugs.
Identify scaling limitations in the current technology stack (TypeScript, AWS, Temporal, Windows Server, RDS) and propose solutions.
Ensure the engineering team has good development environments and test suites.
Enforce coding practices.
Unblock the engineering team.
Ensure infrastructure scales to process more than the current $10 billion of therapies.