Senior Software Engineer

Koda's platform serves real patients at some of the largest health systems and payers in the country, and it has to work during clinical hours. Because we handle PHI, only a small number of US-based engineers can hold production access — and today, that means production issues compete with everything else those engineers do. This role exists to change that: you become the engineer our production systems can count on during US business hours.

We build fast, with AI as a core part of how we ship. That speed is a real advantage, and it creates a real need: someone with strong judgment who hardens what ships, catches problems before patients notice, and jumps on production issues as they come up. The engineer in this role protects the reliability that our health system partners are trusting us with — which makes this one of the highest-leverage seats in the company.

What You’ll Own

Production Health During US Hours

• Serve as first responder for application-level production issues during US business hours, owning them from investigation through the shipped fix
• Share production coverage with the CTO and our infrastructure engineer — app issues are your lane, platform issues are theirs, and you back each other up (no solo always-on pager)
• Keep releases moving: branch syncs, release shepherding, and a CI pipeline healthy enough that shipping stays boring

Reliability & Hardening

• Make fast-shipped, often AI-assisted code production-grade: error handling, retries, edge cases, test coverage, and security review for anything touching PHI
• Review, harden, and land agent-authored PRs, verifying them against how the system actually behaves before they merge
• Build the observability that catches problems before customers do: error reporting, synthetic canaries for critical journeys, and alerting that turns recurring incidents into runbooks and permanent fixes
• Keep the system simple as it grows — refactoring, simplifying, and deleting are as valued here as adding

Data Integrity

• Own the health of our data layer: schemas, indexes, and consistency across related collections
• Design and run the migrations and backfills that clean up inconsistent historical data safely, understanding the blast radius before anything ships
• Add the write-time guardrails that keep data problems from recurring

ADT Ingestion

• Help productionize our ADT (HL7v2) ingestion pipeline: drive it to production, own its monitoring and alerting, and investigate when a hospital feed misbehaves
• Trace issues from feed to patient record and fix the right layer — you don't need HL7 experience coming in; Claude Code and our docs cover protocol details, and your application and data judgment does the rest
• Partner with our infrastructure engineer, who owns the platform side (networking, queues, compliance), while you own the application and data side

How You’ll Work

You'll work directly with the CTO on a small, fast-moving team with broad ownership. Production ownership is the job on day one; as you learn context in the codebase, you'll take features from design to production and grow into full feature ownership. You'll use Claude Code and AI agents as your default tools — pointing them at production issues, verifying what they find, and shipping the fix yourself. This is a remote role with set hours of 9:00 AM to 5:00 PM ET, Monday to Friday, with reasonable flexibility around the edges; consistent daytime availability for production support during US clinical hours is a firm requirement.

• A seasoned application engineer (roughly 5+ years, or equivalent demonstrated judgment) with real production experience: you've debugged live systems under pressure and take pride in operating what you run

• Strong TypeScript across React (frontend) and Node (backend), and comfortable in a GraphQL API and a document database (we use Apollo and MongoDB; comparable experience translates)
• Real database literacy: you can read a schema, reason about indexes and query performance, and write a safe migration or backfill
• Comfortable debugging event-driven systems — queues, retries, redelivery, idempotency — where the failure modes matter more than the wiring
• You already build with AI in your daily loop (Claude Code or similar) and pair it with the discipline to review generated code against domain and security invariants — you don't ship what you can't defend
• Security-conscious by habit: you think about what could go wrong with PHI before it does
• Strong written communication: incident write-ups, runbooks, and decision rationale
• US-based with US work authorization, and able to hold production and PHI-scoped access (background check and BAA as required)

Nice to Have

• Healthcare experience: HL7v2/ADT feeds, SMART on FHIR, Redox, Epic/Cerner, or working under HIPAA or SOC 2 is a strong plus, not a gate
• Observability experience: you've built the alerting or canaries that caught a problem before a customer did
• Basic cloud familiarity: reading logs and metrics, understanding what a queue or CDN is doing to your request (deep AWS expertise not required for our infrastructure engineer owns the platform)
• AWS and infrastructure-as-code experience (we use CDK)
• Real-time and telephony integrations (Twilio, Aircall), or production AI/LLM systems
• Early-stage startup experience where ownership is broad and ambiguity is the norm

Growth & Trajectory

In your first 6-12 months, success looks like production incidents getting rarer and quieter: observability that catches issues before customers report them, a hardened application layer, and the ADT pipeline running in production without drama. From there, your scope grows from operating the system to shaping it. That means taking features from design through production and setting the standards for how we build, ship, and operate with AI. Clear paths exist toward Staff/Principal IC depth or engineering management as the team grows.