We are looking for a Middle+ / Senior Python Developer / AI Engineer for a project on automating the processing and reconciliation of financial documents.

The project involves developing a service that will receive reconciliation statements and internal extracts, extract data from them, match documents, identify discrepancies, and generate a clear result for the user.

Responsibilities

• Parse and process documents in Excel, PDF, and scan formats.
• Extract tabular data from documents.
• Normalize dates, document numbers, amounts, and other fields.
• Design and implement data matching logic.
• Identify discrepancies between two sources.
• Handle complex cases: different periods, differing numbers, partial matches, and other ambiguous situations.
• Integrate OCR and Document AI tools.
• Use LLMs, if necessary, to interpret and generate human-readable explanations.
• Develop the backend/API for the service.
• Participate in the MVP architecture design.
• Work with real anonymized financial documents and test the solution's quality on historical data.

What's important to us

• Commercial Python development experience of 3+ years.
• Strong command of Python.
• Experience with pandas and tabular data.
• Experience processing Excel / CSV.
• Experience with PDF parsing.
• Experience developing backend services and REST APIs.
• Understanding of data matching, normalization, and cleaning algorithms.
• Ability to design a solution independently, not just work from a ready-made technical specification.
• Experience with Git.

Will be a big plus

• Experience with OCR: Tesseract, PaddleOCR, or similar.
• Experience with PyMuPDF, pdfplumber, Camelot, Tabula.
• Experience with fuzzy matching.
• Experience with Document AI / Intelligent Document Processing.
• Integration of LLM APIs: OpenAI, Azure OpenAI, YandexGPT, or similar.
• Experience developing reconciliation / data matching systems.
• Experience with financial, accounting, or other structured documents.
• Experience with On-Premise solutions.
• Experience building solutions that combine deterministic business logic and AI.

What is not required here

We are not looking for a classic Data Scientist and do not plan to train our own ML models from scratch.

What's more important in the project:

• strong applied Python development;
• working with documents and data;
• building reliable matching logic;
• understanding where AI is truly needed and where it's better to use conventional algorithms.

About the project

At the first stage, we plan to develop an MVP without deep integration with internal accounting systems.

Basic scenario:

1. User uploads a reconciliation statement.
2. Uploads an internal extract.
3. The system extracts and normalizes the data.
4. Performs matching.
5. Shows matches and discrepancies.
6. Generates a structured result for the user.

The expected outcome is a working prototype that can be tested on real historical data.

Work format

• Remote work.
• Project-based employment / staff augmentation.
• Full or partial workload possible – discussed individually.
• The project starts with the MVP stage, with the possibility of continuation.
• Work within a small project team.

Expected level

Middle+ / Senior

We need a specialist who can independently understand the subject area, propose a technical solution, and bring it to a working result.

Will be an advantage in your application

If you have experience with similar tasks, please attach a brief description:

document processing; / OCR; / reconciliation / matching; / financial data processing; / Document AI; / LLM integrations.

Also, please specify:

• current availability;
• possible workload in hours per month;
• hourly rate;
• work format: Sole Proprietorship / self-employed / other.