The applied machine learning team at the company builds production AI agents for accounting and finance work.
You take a project from first design and experiments through to deployment, with the aim of making agent systems stronger, more dependable, and easier to measure.
Responsibilities
Build multi-agent setups that automate intricate accounting processes.
Set how far agents can act on their own, which tools they call, how they recover when something fails, and how success is measured.
Create evaluation pipelines, reference tasks, metrics, and experiments that test how reliable models and agents are.
Add monitoring that catches regressions, helps you study errors, and guides ongoing improvement.
Choose which model handles each request and tune routing against latency, cost, and accuracy limits.
Build the prompt, context, memory, retrieval, and indexing layers that let agents reason across tasks.
Turn unstructured documents into structured data and add validation guardrails around it.
Plan projects, record technical decisions, and share progress and lessons with teammates.
Requirements
4 to 12 years of machine learning engineering experience, including applied ML work in industry.
You have shipped end-to-end LLM-based agent applications, covering model orchestration, benchmarking, and testing of behavior and reliability.
Strong Python skills and solid knowledge of LLM and transformer-based systems. Postgres experience is relevant.
You have run structured ML experiments and built evaluation tooling, then improved systems based on measured results.
You have built machine learning products at a quick-moving startup or in a similarly rigorous environment.
Solid technical foundations and clear communication, including explaining complex ideas simply.
An interest in AI and how it affects accounting, finance, and the wider economy.
Conditions
Base salary ranges from $175,000 to $350,000 a year.