Member of Technical Staff

About the role

You will own the design and implementation of core agent infrastructure that operates for hours on end, performing end-to-end work for some of the largest accounting firms in the world. You will own the context ontology that defines how agents perceive and act within complex financial workflows, and you will own the data pipelines that hydrate and update agent state in real time. You will own the eval frameworks that measure agent quality and correctness, and you will own the abstractions that let product teams ship new agent capabilities quickly. You will own experiments that stress how agents handle long horizons, nondeterminism, and partial observability in production. You will own contributions across product engineering, agent engineering, and platform infrastructure as your team's quarterly objectives shift. You will own the technical narrative for how we build reliable, cognate-like AI systems that act as coworkers rather than simple tools. You will own communication with internal and external partners to align on requirements, constraints, and outcomes for high-stakes accounting scenarios.

Key facts

What you'll do

Investigate and prototype context engineering solutions that let agents maintain state and intent across multi-hour accounting workflows.

Design and implement data pipelines that reliably hydrate agent state from heterogeneous sources while preserving lineage and auditability.

Build and maintain eval harnesses that combine automated checks with human judgment to assess agent outputs on real accounting tasks.

Instrument agent trajectories to extract dense signal from long-running sessions and attribute outcomes to specific reasoning steps.

Define and evolve the tool interface abstractions that balance mechanical efficiency with the ability to handle accounting edge cases.

Refactor and optimize our monorepo to improve ergonomics for agent code generation, verification, and review workflows.

Collaborate with the Atlas team to build internal agents and knowledge systems that automate recruiting, sales, and engineering processes.

Partner with product engineers to translate accounting domain workflows into agent-capable architectures that feel like working with a nondeterministic coworker.

Lead experiments that test how users interact with agents who operate for hours, and iterate on interfaces that keep users comfortable with non-deterministic systems.

Own the technical roadmap for agent platform components, including harness engineering, eval systems, and production infrastructure.

Contribute to the design of systems that let the boundaries between frontend, backend, infrastructure, and ML engineering dissolve where it benefits agents.

Champion engineering practices that allow small, cross-functional pods to rapidly reconfigure around new objectives and emerging problem spaces.

Drive the implementation of context layers that encode accounting rules, constraints, and data relationships for use by autonomous agents.

Track and communicate metrics for agent reliability, correctness, and throughput as the team scales and the scope of autonomy expands.

Requirements

You have a Bachelor's or Master's degree in Computer Science, Engineering, or a related quantitative field, or equivalent practical experience.

You have 5+ years of professional software engineering experience, with a strong track record of delivering complex systems in production.

You are fluent in at least one systems-level programming language such as Rust, C++, or Java, and you can learn new languages and paradigms quickly.

You have hands-on experience building and operating distributed systems, including services, databases, and message-driven architectures.

You have experience designing and implementing observability, monitoring, and alerting for high-throughput, low-latency systems.

You are comfortable reasoning about correctness, concurrency, and resource utilization in concurrent and distributed programs.

You have experience building and maintaining data pipelines and ETL workflows that handle large volumes of structured and unstructured data.

You have experience working with machine learning tooling and models, and you understand the tradeoffs between inference latency, throughput, and accuracy in production.

You are comfortable making technical decisions in the absence of complete information, and you can decompose novel problems using first principles and systems thinking.

You are comfortable working with nondeterministic systems and you view agents as coworkers rather than deterministic scripts.

You are able to communicate clearly with both technical and non-technical stakeholders, including product managers, domain experts, and executives.

You are comfortable moving quickly in a rapidly changing environment where priorities and team structures are updated quarterly.

You are legally authorized to work in the United States without sponsorship, and you are willing to relocate to our New York office.

You are passionate about building production-grade AI systems that have real economic impact at scale.

Nice to have

Experience building and deploying agentic systems or complex ML-driven workflows in production.

Deep experience with cloud infrastructure, container orchestration, and service meshes in production environments.

Contributions to open source projects related to agents, data pipelines, or distributed systems.

Experience with accounting, finance, or compliance systems and the unique constraints of regulated data.

Experience building eval frameworks that combine automated metrics with human judgment.

Experience with monorepo tooling, build systems, and developer experience improvements.

Practical notes

Work location is restricted to our New York Office.

This is a full-time role.

We do not specify compensation details on this page.

No visa sponsorship is available for this role.

There are no published application deadlines; interested candidates should express interest in alignment with our current hiring cycles.