AI Financial Data Processing

I’m building a new AI-driven pipeline that ingests raw financial data, cleans and normalises it, then runs machine-learning routines so the results can be consumed by dashboards, reports, or downstream models. The exact data source—stocks, crypto, or transactional feeds—is still being finalised, so the solution must stay modular enough to swap connectors without large rewrites.

I expect the work to centre on Python with pandas, NumPy, scikit-learn (or similar), and a well-structured ETL workflow orchestrated by notebooks or a lightweight API layer. Good documentation and clean, reproducible code are essential; once delivered, my in-house team must be able to extend the models or plug in new data streams without your help.

Deliverables
• A fully functioning preprocessing and feature-engineering module (cleansing, deduplication, outlier handling, enrichment)
• At least one end-to-end example notebook or script that pulls a sample data set, runs the pipeline, trains a basic model, and outputs results ready for visualisation or reporting
• Clear setup instructions plus inline comments so new analysts can follow every step

Acceptance criteria
1. I can point the pipeline at a different CSV source and rerun it without errors.
2. An inexperienced team member can reproduce the example notebook on a fresh machine using only the README.
3. Code quality meets PEP 8 standards and every transformation is traceable for audit purposes.