A complete application in a minute
Tailored resume and cover letter, ready to send.
Codat is seeking an engineer to build and run the data side of an AI assistant used by bankers. You will design and operate data pipelines that clean data from accounting and payment systems, and you will implement shared definitions for terms like revenue to ensure consistency.
As the data steward, you will ship production-grade code in Python or C#, stay responsible for a live system after launch, and ensure safety checks so the numbers are right every time.
The role is for an engineer who builds and runs the data side of an AI assistant for bankers. That means pipelines that clean data from accounting and payment systems, shared definitions for terms like revenue, and checks that catch bad numbers. A strong candidate has shipped tested production code in Python, C# or similar, is sharp with SQL, and has stayed responsible for a live system after launch.
AI agents handle repetitive tasks, and the assistant answers questions straight from a client's books, so the numbers behind it have to be right every time. The engineer's work shifts toward designing trusted data, clear definitions and tools that agents can call safely. Judgment stays human, such as noticing when a requested number can't be supported and explaining why.
The work puts data in front of a language model through MCP servers and APIs, so the experience gained is in building and testing tools that agents rely on. Someone can prepare by building a small MCP server over a messy dataset and adding checks that show when its answers are wrong. Working with banking and accounting data adds a real setting for learning how AI answers get verified.
This is our own read, written from the employer's posting and the company's Spotlight. The employer's posting is the source of truth for the role.
Codat provides an advisory intelligence platform connecting small business financial and accounting data for commercial banking teams, according to the company.