Windows Systems Engineer

Windows Systems Engineer

About Trellis

Trellis builds and deploys computer-use AI agents that get patients access to life-saving medicine. Our agents process billions of dollars worth of therapies a year, automating document intake, prior authorizations, and appeals at scale — classifying medical referrals, understanding chart notes, and automating contract and reimbursement search to give patients accurate coverage determinations and cost responsibility. The result is streamlined operations and faster care for patients in all fifty states.

We're a spinout from the Stanford AI Lab, backed by leading investors including Y Combinator, General Catalyst, and Telesoft Partners, alongside executives at Google and Salesforce.

The Role

We're hiring a Windows Systems Engineer to own the Windows Server and Remote Desktop Services (RDS) infrastructure that powers our fleet of computer-use agents. This is a deeply hands-on systems engineering role: you'll keep the session hosts, service accounts, scheduled tasks, and underlying Windows environments stable, performant, and recoverable — at a scale where downtime has real consequences.

The Scale

The infrastructure you'll own operates at serious scale and directly supports patient care:

• Thousands of computer-use agents running concurrently, 24/7

• 20M+ prescriptions processed every year

• 400 clinics live across all 50 states

• $15B+ in medications dispensed flowing through the systems you keep running

When a session host hangs or an agent stalls, it isn't an abstract SLA breach — it's a clinic waiting on a prescription. Reliability here is the job.

Key Responsibilities

Windows Server Operations

• Administer and maintain a fleet of Windows Server hosts: patching, reboots, and upgrades with minimal disruption

• Diagnose and remediate runaway processes, memory leaks, high CPU, and zombie sessions on live hosts

• Manage scheduled tasks for agent startup, watchdogs, and automated recovery

• Configure networking, service wrappers, and per-agent tunneling

• Own the identity and access plumbing — service accounts, permissions, and Group Policy / Active Directory where it touches the fleet

RDS & Session Management

• Administer RDS Session Hosts: session limits, autologon, persistence, and recovery of dropped or frozen sessions

• Manage RDS licensing (CALs, grace periods) and troubleshoot licensing-related failures

• Handle loopback/local RDP architectures and use session shadowing to debug live agents

Fleet Provisioning, Scaling & Lifecycle

• Provision and decommission agents on demand, including service accounts, sessions, and per-agent environments

• Manage capacity, rebalance load across hosts, and scale session-host capacity up or down to match demand

• Roll out upgrades safely with staged/canary rollouts and clear rollback plans

• Maintain golden images so hosts and agents come up consistently and predictably