About the Role

Join an applied machine learning team building AI systems for accounting and finance workflows. You will own projects from initial design through production, developing agent systems that reason, plan, and improve through evaluation.

What You'll Do

Design and evolve multi-agent architectures that automate real-world accounting workflows.

Build safe agent behaviors, including autonomy boundaries, tool use, memory, and fallback strategies.

Route, evaluate, and optimize models for latency, cost, and accuracy.

Create automated offline and online evaluation pipelines, defining tasks, labeling strategies, and measurable performance criteria.

Instrument systems to detect regressions, categorize errors, and guide continuous improvement.

Build prompt, retrieval, indexing, and document-processing systems that provide agents with relevant context.

Write concise specifications, build and test systems end to end, and communicate project progress and findings.

What We're Looking For

Four to twelve years of machine learning engineering experience, including applied industry work on machine learning products and models.

Hands-on experience delivering end-to-end LLM-based agent applications, including benchmarking, model orchestration, and evaluations of agent behavior and reliability.

Strong Python skills and deep familiarity with LLM and transformer-based systems; experience with Postgres is relevant.

Experience designing structured ML experiments, building evaluation infrastructure, and iterating based on measurable results.

Background in a fast-paced startup, quantitative finance, or a highly technical organization, with most career experience outside large technology companies.

Strong technical fundamentals, clear communication, and the ability to explain complex concepts simply.

Interest in applying AI to accounting, finance, and broader economic workflows.

Compensation & Benefits

Base salary: $175,000 to $350,000 annually. Visa sponsorship is available.

Location

On-site in New York, NY.