About the role
Pennylane is a French fintech building an accounting OS for SMEs and their accountants, now used by more than a million businesses. The Embedded AI team owns Copilot and Autopilot: invoice parsing, document classification, matching, scoring, and bookkeeping suggestions that have to hold up in production.
You take those problems from framing through data and labeling, training, evaluation, deploy, and monitoring. The bar is product metrics such as automation coverage, human correction rate, latency, and cost. The stack they name is Python, PyTorch, PySpark, Redshift, Airflow, and AWS SageMaker. They want 5-8 years of production ML. English is enough; French is not required.
Highlights
- Europe remote within CET ±2h; Italy is listed among hiring countries
- 5-8 years shipping ML systems, not research handoffs
- Python, PyTorch, Spark, Redshift, Airflow, SageMaker
- Copilot and Autopilot on real accounting workflows
This recap is dataskew's editorial summary, not the company's copy.