Staff Software Engineer, Infrastructure

Wisdom • New York, NY

1d • Remote

About The Position

The roadmap isn't handed to you here. You'll help write it — and you'll be the reason it stays up.
As a Staff Software Engineer focused on Infrastructure at Wisdom, you'll set the technical direction for reliability across the company — and own the systems behind the systems: the deploy pipeline, the observability, the capacity controls, and the failure-handling that decide whether our agentic billing infrastructure quietly does its job or pages someone at 2am. This is a force-multiplier role on a small, high-trust team. Your job isn't just to fix what breaks; it's to make the whole organization operate at a higher reliability bar — to build the practices, the guardrails, and the instincts that mean fewer things break in the first place, and the team can handle the ones that do without you in the room.
Wisdom's stack is TypeScript, Node.js, React, Postgres, and AWS, with LLM-driven agents (Mastra, Anthropic) making high-stakes billing decisions in production. The problems we're solving — keeping inconsistent insurance integrations alive, making AI pipelines fail safe instead of failing loud, running HIPAA-compliant infrastructure that genuinely can't go down — are legitimately hard. We'd rather have someone energized by making things not break than someone who merely tolerates being paged when they do.
In your first year, you'll have defined what reliability means at Wisdom and built the function to deliver it: a real observability and SLO practice, an incident process that runs without heroics, agentic pipelines that degrade gracefully instead of taking prod down with them, and a team that's measurably better at operating production because of how you've raised the bar. This is a fully remote role reporting directly to the Head of Engineering.

Requirements

• 8+ years running production systems, with a track record of operating at staff/principal scope — you've owned reliability for systems where downtime had real consequences and left them measurably better
• You've operated at scale under pressure — services that had to stay up, incidents you led to resolution, and reliability practices you established that outlived your tenure and changed how teams worked
• You multiply the people around you — your impact shows up in what others ship reliably, not only in what you touch directly. You've set standards, mentored engineers, and driven technical decisions across teams without needing the authority to mandate them
• Deep AWS (or GCP) experience — you've deployed, operated, and debugged distributed services in production, and can reason from first principles when the runbook runs out
• Strong with infrastructure as code (Terraform), containers and orchestration (ECS/Kubernetes), and CI/CD — the deploy path is yours to make boring
• Hands-on production experience operating at least one major LLM API — OpenAI, Anthropic (Claude), or Google Vertex AI — with a focus on the operational reality: rate limits, retries, latency, cost, and what happens when the model misbehaves in a live system
• Strong command of TypeScript/JavaScript — you can read and fix the application code, not just the infra around it; Python or Go a plus
• Deep experience with relational databases — connection management, query performance, and reasoning about data integrity under load
• You default to ownership and move toward the pager, not away from it
• You're direct, intellectually honest, and collaborative — you surface bad news early, change your mind when the evidence warrants it, and write the postmortem that makes the whole team sharper

Nice To Haves

• Experience operating LLM / agentic systems in production — or with frameworks like Mastra, LangChain, LlamaIndex, or CrewAI — where reliability, cost, and latency were yours to define and manage
• Working knowledge of HIPAA compliance and what it means to run infrastructure responsibly in a healthcare context
• Experience at a Series A or early-stage startup where you built the reliability function from scratch rather than inheriting one

Responsibilities

• Set the reliability strategy for the platform — SLOs, error budgets, and the operating standards for services that bill real money for real practices, and the technical roadmap to get us there
• Own observability end-to-end — tracing, metrics, logging, and alerting (Datadog) that surfaces problems before users do, not after — and make it the default so any engineer can lead an incident, not just the person who wrote the code
• Define how we operate AI-powered agentic workflows at scale — retries and backpressure, idempotency, graceful degradation, and capacity controls for LLM-driven pipelines. The failure modes here are new (batch blowups, stream drops, runaway cost, model misbehavior); you'll be writing the playbook the rest of the industry hasn't written yet, and setting the patterns the team builds against
• Harden the integration surface with dental insurance carriers and practice management systems (Dentrix, Eaglesoft) — poorly documented, inconsistent, and the first thing to buckle under load
• Own deploy and release engineering — fast, safe, reversible deploys; infrastructure as code (Terraform); and the unglamorous discipline that lets a Series A ship many times a day without breaking things
• Build the incident practice, not just lead incidents — the on-call rotation, the runbooks, the blameless post-incident culture, and the follow-up discipline that turns outages into permanent fixes the whole team owns
• Raise the bar through others — set technical standards via code review, architecture guidance, and documentation that actually gets used, and level up how the entire engineering team reasons about reliability
• Take on the ambiguous, undefined, company-level reliability problems and drive them to resolution without waiting for permission or a perfect brief

What This Job Offers

• Job Type: Full-time
• Career Level: Senior
• Education Level: No Education Listed