Lead Software Engineer - medical billing platform - agentic ai automation

Please note: The the salary range listed is the actually range - please do not ask for an amount above. All candidates must currently reside in the state of Texas. No relocation is offered. No work sponsorship is available.

The Opportunity:

We are building a new billing platform from the ground up and need the senior engineer who owns its core: the engine that moves claims and payments through their lifecycles, the background job system that runs every automated step, and the integrations with payers, clearinghouses and banks. Backend engineering on a money system, TypeScript on PostgreSQL on Azure, working directly with the CTO on a team that stays at two to four engineers.

The type of engineer we are seeking:

A backend engineer who has shipped a transactional system (payments, claims, orders, ledgers, anything where a row posting twice is a real problem) in a typed language, and who has been the one who found out when it broke.

Experience you must have:

• Seven or more years of backend engineering in production, in a typed language: TypeScript or Node.js, C#, Java, Go or Kotlin. TypeScript is what we use. A strong C# or Java engineer who has not written much TypeScript is fine; a Python-only engineer is not.
• A relational database as their primary store for years. PostgreSQL preferred; SQL Server, DB2, Oracle or MySQL fine. They should be able to talk about transactions, locking and a deadlock they debugged.
• Built and ran something with background workers, queues and retries. The tool does not matter (Kafka, RabbitMQ, SQS, Azure Service Bus, Hangfire, Quartz, Sidekiq, Celery, a database table). The years running it do.
• Can explain idempotency and "what happens if the process dies between these two writes" from a system they built.
• Integration work against outside vendors: REST or SOAP APIs, SFTP file drops, fixed-width or EDI files, vendors who change their output without warning.
• Owned production. They found out their job broke before the business did and changed the design so it could not repeat. A formal on-call rotation is evidence, not the requirement.

Questions you will be asked:

• "Tell me about a system you built where the same job must never run twice at the same time. How did you stop it?"
• "Tell me about a production failure you owned where nothing errored and the data was just wrong or missing. What did you change afterward?"
• "What is the most TypeScript you have written?"

The pitch to candidates:

• Greenfield core of a real money system, with an architecture already written down and a CTO who will hand it over rather than hover.
• Small team, no service fleet, no twelve microservices, one database to search when something goes wrong.
• Every automated step earns trust on evidence (shadow, propose, auto). The engineer who builds that gets to decide how far it goes.
• Decisions are written. They will read the ones that exist and write the next ones.
• Billing is unglamorous and sticky. The agencies are the customer, the money is real, and the problems are concrete.