Finance & AI Onboarding Intern

At Cortena, we're building the AI execution layer for finance operations in SMEs. We have paying clients, real workflows running in production, and a product that is evolving fast.

Job description

Why this role exists

New clients rarely arrive with a single clean export. They have multiple partial sources of truth: chart of accounts, creditor/debtor masters, cost centres, BU codes, VAT codes, and often long booking histories with inconsistent descriptions. Your job is to work with each client only as much as needed, infer structure from messy data, and turn it into clear rules and reference data so our AI pre-accounting stays accurate.

You bridge real-world accounting messiness and how Cortena models reference data (masters in spreadsheets / connected accounting systems + guideline documents). You will work directly with our founding team so clients go live smoothly, outputs stay trustworthy, and the product improves from what you see on the ground.

This is not a ticket-chasing support role. You will have real ownership from day one.

What you will do

Client reference data & onboarding (core)

• Own “source of truth” discovery with minimal friction: figure out the smallest set of asks (files, exports, system access) to obtain chart of accounts, creditor/debtor lists, cost centres, BU/VAT context where relevant, and historical postings when masters alone are incomplete.
• Normalize inconsistent inputs: combine spreadsheets, CSVs, PDFs, and accounting exports; resolve naming chaos, duplicates, and missing codes; align creditor/debtor identifiers with booking behaviour where needed.
• Infer structure from booking history: mine long transaction lists to recover stable patterns, deduplicate noise, and use that to validate or complete chart-of-accounts and allocation logic.
• Capture how the business actually operates: identify the client’s main line(s) of business, typical spend, and recurring ambiguous line descriptions; turn that into practical matching rules when the system must choose between accounts, cost centres, or BU codes.
• Translate into Cortena setup: ensure reference data and guidelines are reflected in our tenant configuration, including connected systems, Google Drive reference sheets, and guideline sources the product uses for classification, and flag gaps to the product team before they become client incidents.

AI Quality Assurance

• Review AI-processed invoices and transactions against the reference data and guidelines you helped define; catch systematic misclassifications, not one-off quirks.
• Flag edge cases and misclassifications, document patterns, and feed insights back to the team.
• Document patterns (good and bad), suggest evaluation checks and prompt or flow improvements, and help keep quality measurable as changes are shipped.
• Help develop and maintain evaluation frameworks to measure output quality over time.
• Contribute to prompt testing and improvement as our AI workflows evolve.

Client Communication

• Be a clear, structured point of contact for onboarding data questions, focused on the data and setup thread.
• Turn client reality into short, actionable notes for engineering: what was messy, how you resolved it, and what the product should do better next time.
• Act as a first point of contact for onboarding questions and technical setup issues.
• Translate client feedback into clear product and improvement notes for the team.
• Keep onboarding documentation up to date and easy to follow.

What we are looking for

Must-have

• Genuine interest in AI applied to real finance work — you use AI tools and understand limits, including hallucinations and brittle rules, as well as strengths.
• Solid intuition for bookkeeping and SME finance: invoices, GL accounts, cost allocation, creditors/debtors, and how posting history relates to masters.
• Structured problem-solving: comfortable with large tables, reconciliation, and “good enough” triage when data is incomplete.
• Excellent communication: you can say what is wrong, why it matters, and what should be done — briefly.
• Detail-oriented — you notice when something is off and follow it through.

Nice to have

• Exposure to DATEV, Exact Online, Twinfield or similar; comfort with Excel/CSV and Google Sheets at volume.
• Experience turning ambiguous natural language, such as invoice lines and memo fields, into routing rules or guidelines.
• German language skills.

What you will get

• Direct exposure to founding-team decision making at an early-stage AI startup.
• A front-row seat to how a real AI finance product is built and iterated.
• Hands-on experience with production AI workflows and real client data.
• Flexible hours — 20 hours per week or full-time, depending on your availability.

Cortena is an AI-native finance operations platform for SMEs in Europe. We automate the manual, repetitive work that finance teams deal with every day — invoice processing, pre-accounting, bank reconciliation, and more.

We are a small, focused team. Every person we bring in shapes what we become.