Docflow builds AI agents that run the back office for healthcare providers. Our automation handles the work that buries their teams: document intake, prior authorizations, order entry, and payer portal workflows. Customers start with a paid pilot, see real work getting done within weeks, and expand from there.
We’re a small, scrappy, AI-native founding team, and we’re growing fast. Every customer we sign gets a tool built around how their team works, and there are more of them every month. That’s why we’re hiring you.
The RoleYou’ll own the full experience of building and running a tool with Docflow, from the long-running workflows underneath to the product customers use every day.
A Docflow tool is two things at once. Underneath, it’s a set of long-running workflows that pull in faxes and orders, work through payer portals and the customer’s own systems, wait days for an authorization to come back, and write results where the customer’s team already works. On top, it’s a product: the pages where a customer launches the tool, watches it work, reviews exceptions, sees what it saves them, asks questions, and requests changes. Customers judge us on both, so you’ll build both.
The job spans a tool’s whole life. An operations lead describes a messy process on a kickoff call. A few weeks later, their team is trusting software with prior authorizations. Months after that, a payer redesigns its portal or the customer changes a rule, and they want to know what changed and whether it worked. Every one of those moments is product: how a tool gets configured, how a run explains itself, how an exception tells a person exactly what to do next, and how a change request becomes a verified fix the customer can see.
This is a full-stack role. In one week you might design a workflow that checks back on a pending authorization every six hours for two weeks, trace why a scheduled job stalled overnight, and ship the page that shows a customer why their runs needed human review. You need to be as comfortable with a component’s loading and error states as with a workflow that has been running since last Tuesday. Behind every run is a patient waiting on equipment or care, so a stalled workflow or a confusing exception is never just a bug.
Agents already investigate most failures and write most fixes, and our Platform Engineer builds the system they work in. You’ll build on that system and hold it to the customer’s standard: did the fix solve their problem, and can they tell? And whenever a customer has to ask us for something, you’ll ask whether the product should have handled it. That instinct is how customers got self-serve credentials, schedules, and billing, and run pages that explain themselves.
You’ll work alongside the account executive who owns each customer relationship. They stay in the lead with the customer, and you make what the customer needs real.
What You’ll DoBuild new customer tools end to end alongside the account executive: the workflows, the integrations with payer portals and the systems customers run their business on, and everything a customer needs to launch
Operate and maintain long-running Temporal workflows in production: schedules, multi-day check-backs, retries, and exception routing that sends the right work to the right person
Build the product customers use every day: runs and jobs, exception review, analytics that show what the automation is worth, quality reports, credentials and schedules, and the assistant they ask questions
Own the path from a customer’s change request to a verified fix they can see, from how the request gets filed and approved to how the result shows up on their changelog
Make every tool explain itself in the customer’s language, with run messages, step descriptions, and write-backs that name the systems and people their team knows
Turn repeated customer requests into self-serve product, so the next customer never has to ask
Spot patterns across customers and tools, and feed them back into the platform, the agents, and how we build the next tool
Learn our customers’ industry well enough to argue with how they describe their own process
Must have:
Strong professional experience shipping production web applications end to end. We care more about what you’ve built than a years-of-experience number, though in practice this usually means several years in the field.
Real comfort across the full stack: relational databases and type-safe data access, backend services, and modern component-based UI with thoughtful loading, error, and optimistic states.
Experience building and operating long-running or asynchronous systems, such as workflow engines, job queues, or event-driven services. Retries, idempotency, timeouts, and failures that surface days later are things you think about by default.
Product instinct. You think about who’s on the other side of the screen, what they’re trying to get done, and what they’ll misread, and you sweat the details that make software trustworthy.
Strong TypeScript, or strong experience in a comparable typed language plus the appetite to get to strong TypeScript quickly.
You build with AI coding agents such as Claude Code, Codex, or Cursor every day.
Genuine interest in complicated real-world processes. Prior authorization rules, billing codes, payer-by-payer exceptions that exist for reasons nobody ever wrote down. If a twenty-step workflow that nobody has documented sounds tedious rather than interesting, this will be a frustrating job.
Comfort in a company of four to six people, where the job is whatever it takes and you’ll largely define your own benchmarks as you go.
Willingness to work in-office five days a week, and to be around some evenings and weekends when something critical breaks.
Preferred:
Customer-facing engineering experience, such as forward deployed, solutions, or implementation work where you also wrote production code
Experience with durable workflow orchestration such as Temporal
Experience designing software for operations teams, such as dashboards, review queues, and self-serve configuration
Experience in healthcare, revenue cycle, insurance, or another domain with dense rules you had to learn from scratch
Experience with browser automation, document extraction, or integrating with systems like EHRs and CRMs
A strictly typed, functional style: discriminated unions, exhaustive handling, and state modeled as explicit values
Based in Los Angeles, or ready to be
We’ll teach you the industry. We can’t teach you to find it interesting.
Compensation and BenefitsBase salary of $150,000 to $185,000 depending on experience
Founding-team equity
Health insurance
Skills Required
- Strong professional experience shipping production web applications end to end
- Experience with relational databases and type-safe data access
- Backend services development experience
- Modern component-based UI development, including thoughtful loading, error, and optimistic states
- Experience building and operating long-running or asynchronous systems, such as workflow engines, job queues, or event-driven services
- Understanding of retries, idempotency, timeouts, and delayed failure handling
- Strong product instinct and attention to trustworthy user experiences
- Strong TypeScript or strong experience in a comparable typed language with willingness to learn TypeScript quickly
- Daily experience building with AI coding agents such as Claude Code, Codex, or Cursor
- Genuine interest in complicated real-world processes and undocumented workflows
- Comfort working in a four-to-six-person company with broad responsibilities
- Willingness to work in-office five days per week and be available some evenings and weekends for critical issues
- Customer-facing engineering experience, such as forward deployed, solutions, or implementation work involving production code
- Experience with durable workflow orchestration such as Temporal
- Experience designing software for operations teams, including dashboards, review queues, and self-serve configuration
- Experience in healthcare, revenue cycle, insurance, or another domain with dense business rules
- Experience with browser automation, document extraction, or EHR and CRM integrations
- Experience with a strictly typed functional programming style, including discriminated unions and exhaustive handling
- Based in Los Angeles or willing to relocate there
What We Do
Docflow builds AI agents that run the back office for specialty healthcare providers — the organizations that get equipment, supplies, and ongoing care to patients outside the hospital. The work that buries these teams is not clinical. It is document intake, prior authorizations, order entry, eligibility and benefits checks, and endless hours inside payer portals. A single patient order can mean a faxed referral to read and classify, a prior authorization to submit to one of a dozen portals, a follow-up a week later to find out whether it was approved, and a stack of updates typed back into the provider's system of record. Multiply that by thousands of orders a month and you have entire departments doing work no one went into healthcare to do. Our agents do that work. They log into the same systems a person would — the provider's system of record and dozens of payer and utilization-management portals — read the documents, fill the forms, and write the results back. Every action is recorded: a screenshot and a full session trace for every step, so a provider's team can audit exactly what was done on every order. When an agent hits something it should not decide alone, it stops and asks a human instead of guessing. We are HIPAA compliant and handle PHI for every customer we serve. Customers start with a paid pilot on a single workflow, see real orders getting processed within weeks, and expand from there. Most of our growth comes from customers handing us their next workflow. We are a small founding team based in North Hollywood, Los Angeles. We are revenue-funded, our recurring revenue quadrupled over the last five months, and our pipeline is bigger than we can currently work.
Why Work With Us
We are revenue-funded and small enough that your work is visible the day you do it. Recurring revenue quadrupled in five months and our pipeline is bigger than we can work. You'll work directly with the founders, own a function outright rather than a slice of one, and watch AI do real work in a corner of healthcare nobody has automated yet.
Gallery
.png)
