Senior Software Engineer, Agentic Engineering
About Fathom
Fathom is on a mission to eliminate the billions of dollars of administrative waste in US healthcare. We're starting with one of its most expensive, labor-intensive workflows: medical coding. Using AI, we automate the translation of clinical notes into the billing codes used for provider reimbursement—a process that costs US hospitals $15B+ annually, plus tens of billions more in errors and denied claims.
The result is healthcare that spends less time and money on administration and more on what actually matters: patients. KLAS Research—the industry standard for healthcare IT ratings—named us the #1 emerging technology for reducing the cost of care. Many of the nation's largest health systems, health plans, and physician groups rely on us to do it.
We're a Series B company backed by Lightspeed, Founders Fund, and CVS Health, taking on hard problems at the frontier of AI and healthcare, where getting it right has real consequences. We're scaling fast and looking for exceptional people who want their work to matter.
About the role
We are hiring a Senior Software Engineer, Agentic Engineering to build the products and infrastructure at the core of that system. You will build and operate the systems that let fleets of coding agents ship real work safely and fast: the pipelines, evaluation harnesses, and the verification and permission rails that turn raw agent output into trustworthy, production-grade software at a scale no individual coder reaches alone.
The bottleneck is no longer writing code: it is reviewing and merging it while managing risk at scale. The skill we are hiring for is judgment at scale: deciding what needs line-by-line review versus empirical verification, where to add rails (tests, egress and permissions controls), and which problems are worth solving first.
We care more about trajectory than tenure: we're looking for engineers who take on scope and impact fast and who grow into technical leadership at Fathom quickly. If that's you, apply, whether you'd call yourself senior yet or not.
This is an in-office role based in New York City (Manhattan Financial District). We build in person, alongside our leads and engineering team on-site.
What you'll do
You'll work agent-first across the stack: building the systems agents operate, and the rails that make their output safe to ship. A sample:
Build the systems that let agents extend our medical-coding engine: the rules, models, and pipelines that turn raw clinical documents (EHR exports, HL7/Cloverleaf feeds, scanned documents) into accurate, reimbursable billing codes at production scale.
Scale data pipelines, not just build them: ingest and normalize high-volume, messy medical data from dozens of disparate sources, and keep those pipelines fast, reliable, and cost-efficient as volume grows.
Build the guardrails for an agent-first codebase: as coding agents take on more of how we build and ship, controlling what they can read, write, and reach (egress controls, least-privilege access, permission rails) is first-class engineering, not an afterthought.
Advance agentic CI/CD, testing, and verification: when agents write a large share of the code, the hard problems move to shipping it safely. You build the testing, CI, and empirical-verification systems that let agent-generated changes merge fast without breaking production. These are open industry problems we confront head-on.
You may be a good fit if you have
Demonstrated leverage from coding agents: you have strong opinions about workflows and rails that let agents ship real work at scale, and how those workflows and rails likely need to evolve as coding agents and harnesses improve, and you reason clearly about risk (what needs review versus what can be verified empirically).
Strong command of algorithms, data structures, and systems design.
3+ years of software engineering experience in a production setting.
A generalist's instinct: you reach for whatever the problem needs and take ownership quickly, rather than staying in a single specialty lane.
Experience scaling data pipelines, making them fast, reliable, and cost-efficient as data volume and source count grow.
Solid command of databases and large-scale data processing: cloud data warehouses (e.g., BigQuery), distributed processing (Spark, Beam/Dataflow), and pipeline/orchestration tooling (dbt, Airflow), plus the judgment to pick the right tool for the job.
The creative and analytic range to design a system that pulls together, trains, and tests dozens of data sources.
A growth mindset and the willingness to operate within, and improve, how a fast-moving team works.
A desire to build in person at our New York City office.
Bonus points if you have
Experience building, evaluating, and shipping machine-learning and LLM-powered systems: especially structured prediction or extraction over messy real-world text.
Hands-on experience building or scaling cloud compute and distributed training or data-processing infrastructure.
Expertise wrangling healthcare data and/or HIPAA.
Experience building the pipelines that turn expert human labels into model training and evaluation data at scale.
What we offer
Salary: $195,000 USD - $250,000 USD
Company Equity
PTO and Uncapped Sick Days
Medical/Dental/Vision Coverage
401k Matching