AI/ML Intern: Latent Reasoning for Tabular Models

AVP; Paris; Vor Ort; EUR 18.000 - 22.000; Vollzeit

Benefits dieser Stelle

Research experience

Collaborative team

Career growth

Publications

Networking

Sustainable AI

Zusammenfassung

Ekimetrics in Paris invites a final-year Master’s or Engineering student to join its Innovation Department for a 6-month internship focused on Latent Reasoning for Tabular Foundation Models. You will work with senior researchers to design and evaluate latent reasoning mechanisms, implement them on open-source TFMs, and study interpretability methods, contributing to production-grade data science workflows.

The role offers hands-on research, collaboration with PhDs, and pathways to a full-time

Qualifikationen

• Currently in the final year of a Master’s or Engineering degree with strong foundations in ML, statistics, and programming.
• Good grasp of model validation: train/test splits, data leakage, baselines.
• Strong analytical skills and ability to work autonomously on open-ended research problems.
• Knowledge of transformers and language models; familiarity with latent reasoning or recurrent-depth models is a plus.
• Strong Python and PyTorch skills, plus pandas/polars/scikit-learn in standard data science toolkit.

Aufgaben

• Review latent reasoning literature (NLP) and tabular foundation models.
• Identify and adapt latent reasoning mechanisms to TFM architectures.
• Implement them on open-source TFMs and measure their effect on accuracy, calibration and robustness against strong baselines.
• Apply interpretability methods to characterize the computation performed during latent reasoning.
• Build an evaluation protocol on diverse tabular datasets, including realistic business cases.
• Document results and contribute to scientific publications.

Kenntnisse

Machine learning

Statistics

Python

PyTorch

Transformers

Data science toolkit

Pandas/Polars

scikit-learn

English

Ausbildung

Master’s degree in ML/CS

Tools

PyTorch

Transformers

Databricks

Jobbeschreibung

Ekimetrics in Paris invites a final-year Master’s or Engineering student to join its Innovation Department for a 6-month internship focused on Latent Reasoning for Tabular Foundation Models. You will work with senior researchers to design and evaluate latent reasoning mechanisms, implement them on open-source TFMs, and study interpretability methods, contributing to production-grade data science workflows.

The role offers hands-on research, collaboration with PhDs, and pathways to a full-time