Founding AI Engineer

Andera • San Francisco, CA

2d

About The Position

Andera is building reliable AI agents for back-office finance, beginning with automating audits for large public companies. The work addresses the scale and instability of financial data, including hundreds of millions of tokens across large spreadsheets with changing formats. Andera aims to turn weeks of manual audit work into minutes of review by building agents that can parse, interpret, and reason over complex data, supported by robust data systems, long-horizon reasoning, and production-grade reliability.

The AI Engineer will help build the core agent systems that power autonomous financial audits. The role involves long-horizon reasoning, large-scale data ingestion, and production-grade agent reliability, with close collaboration with auditors and engineers. It is a high-ownership position spanning architecture through deployment.

Requirements

• Experience building complex agent or workflow-based systems, with the ability to reason about rebuilding systems like Claude-style agents from scratch.
• Understanding of agent planning, tool use, memory, and failure modes.
• Strong full-stack engineering ability and the ability to ship fault-tolerant features end to end.
• Understanding of how code changes interact with the broader system.
• Strong systems engineering fundamentals, including designing concurrent workloads over large datasets.
• Careful consideration of throughput, backpressure, retries, and observability.
• Ability to learn non-engineering domains quickly and translate concepts from domain experts into working systems.

Nice To Haves

• Experience with or adjacent to TypeScript, Python, Temporal, Docker, Terraform, or AWS Fargate.

Responsibilities

• Design and implement complex AI agent systems, including planning, memory, tool use, and recovery.
• Build fault-tolerant, end-to-end features that integrate with the broader codebase.
• Split and execute workloads processing approximately 100GB or more of data efficiently across processes and threads.
• Debug agent failures, edge cases, and production performance bottlenecks.
• Learn audit workflows and implement agent improvements quickly.

What This Job Offers

• Job Type: Full-time
• Career Level: Senior
• Education Level: No Education Listed