AI Backend Engineer

AI Backend Engineer (Self-driving Bookkeeping Architecture)

Location

We’re based in Amsterdam and are open to both in-person and remote candidates. We value time together for fast iteration, but we’re pragmatic and will optimise for the best person for the job.

Employment Type

Thred website

About Thred

Thred

Small businesses run on operational platforms such as POS, scheduling, and inventory systems, but their financial data remains fragmented across banks, documents, and disconnected tools. This makes profitability, liquidity, and VAT and tax obligations hard to manage.

Thred enables these SME platforms to embed accounting, bookkeeping, and tax workflows directly into their product.

If you get these two principles, you get Thred: they shape every product decision we make.

1. 🪢 Embedding Accounting into the SME Platform 🪢

2. 🤖 Self-Driving Bookkeeping with AI Agents 🤖

About the Role

We are hiring an AI Backend Engineer to build the backend systems that power Thred’s AI agents and automation workflows. You will design and ship production workflows that turn operational and financial events into accurate accounting outputs, with clear audit trails and safe exception handling.

This is a hands-on AI agent engineering role focused on reliability and correctness. You will work closely with product and bookkeeping domain experts to automate high-volume bookkeeping tasks, route edge cases for review, and continuously improve system quality through evaluation and feedback loops.

What You Will Do

• Build backend services that run Thred’s AI agents across document intake, transaction categorisation, journal drafting, reconciliation, and month end
• Design event driven pipelines that ingest platform events, bank feeds, and documents and produce draft postings, tasks, and exceptions
• Implement RAG workflows that load business context and jurisdiction specific accounting and VAT rules at runtime
• Add confidence scoring and exception routing so automation is safe and reviewable
• Build evaluation and regression suites with golden datasets, automated checks, and monitoring
• Own production fundamentals including observability, cost and latency controls, versioning, and multi tenant security
• Partner with product and bookkeeping experts to turn requirements into reliable workflows

About You

• You have strong backend engineering fundamentals and can ship and operate production systems.
• You can code in Python fluently as we code all the backend in Python(for now)
• You have built LLM powered workflows, AI agents in production, including tool use, writing MCP servers, structuring outputs from AI models, and failure handling.
• Have worked with vector databases, or Reinforcement learning feedback tools and are eager to learn these tools faster.
• You

  DEEPLY

  care about correctness, auditability, and safe automation, especially in finance and compliance contexts.
• You communicate clearly and work well with domain experts to turn messy real world problems into robust software.

The process

• Intro call to align on role, scope, and ways of working
• Practical exercise focused on an event to accounting output workflow and evaluation
• A deeper technical session on the exercise output.
• Culture-fit call with cofounder & CEO Mark Holleman