Ekimetrics: Data Scientist Internship - Causal Fairness Analysis: From Theory to Practice 🤓
Ekimetrics, a leading data science and AI solutions provider, is seeking a passionate and talented intern to join their Causalité team in the Innovation department.
About Ekimetrics 📊
- 400+ data science experts
- 1000+ diverse projects for 350+ clients
- 5 offices: Paris, Hong Kong, Shanghai, London, and New York
- $1 billion in profits generated for clients since 2006
Our Mission 🌱
We aim to help our clients rethink their business models by aligning economic performance with sustainable goals through data science. To achieve this, we have a diverse team of experts, including Product Managers, Product Designers, Data Architects, Lead Tech, Data Engineers, DevOps Engineers, and Data Scientists.
Join the Data Science Team 🤝
Why are we recruiting?
Our Innovation department collaborates with industry and academic partners to conduct research in AI. We have several PhD experts in generative AI, deep learning, computer vision, time series, explainability, and causality. Two CIFRE theses are ongoing, and two more will start in 2025. Each expert leads a team responsible for testing state-of-the-art algorithms, adapting them to specific business problems, developing new methodologies or algorithms, and ensuring integration into Ekimetrics' industrial ecosystem.
Causalité Team
The Causalité team focuses on a wide range of causal tasks, including causal discovery, causal effect estimation, and counterfactual reasoning, through cutting-edge research in causal inference. We leverage the latest AI and statistical modeling techniques, such as machine learning, deep learning, and Bayesian modeling. By continuously testing new algorithms and auditing Ekimetrics' practices in causal reasoning, the Causalité team refines existing methodologies and develops new ones to address novel problems and expand Ekimetrics' expertise. This allows Ekimetrics to move beyond analyzing correlations and remain at the forefront of data-driven decision-making, empowering clients to utilize causal reasoning.
Internship Objectives: Causal Fairness Analysis: From Theory to Practice
Artificial intelligence (AI) is widely used in businesses and institutions as a decision-making tool in various applications, including sensitive sectors such as healthcare, politics, and finance. While AI systems can provide a competitive advantage, it is crucial to ensure that fundamental human rights are respected. As a result, AI systems are increasingly regulated, particularly in Europe with the AI Act, which requires all high-risk AI systems to assess their compliance and impact on fundamental rights, including the right to non-discrimination.
Fairness analysis in AI systems, commonly referred to as "fairness analysis," is based on two key concepts: "disparate treatment" and "disparate impact." Causal reasoning is a necessary perspective to consider, especially regarding disparate treatment and impact notions, although it is not the only way to derive a fairness analysis. Recently, Plečko and Bareinboim developed a framework for causal fairness analysis (CFA) based on these doctrines and translated them into mathematical language using counterfactual formulations. While the CFA framework is rich and theoretically sound, its practical implementation in real-use cases remains challenging (e.g., uncertainties related to the validity of assumptions and technical complexities in identifying and estimating counterfactuals in finite data regimes).
Therefore, this internship aims to implement and evaluate the CFA framework in real-use cases. The implementation will be done using Double Machine Learning (DML) approaches and integrating sensitivity analysis components. Regarding real-world applications, we will focus on the finance and utility sectors, starting with the COMPAS dataset.
This internship offers an opportunity to engage in cutting-edge AI research and potentially pursue a PhD program or a career as an AI Data Scientist with Ekimetrics.
Your Responsibilities
- Develop knowledge in causal inference and fairness in AI
- Conduct a comprehensive literature review of state-of-the-art models and techniques in causal AI fairness
- Implement and experiment with double machine learning (DML) approaches enhanced by sensitivity analyses
- Collaborate with your supervisor to design, implement, and evaluate prototype models on synthetic and real datasets
- Share findings with internal teams and contribute to developing long-term strategies for responsible AI at Ekimetrics
Your Profile and Skills
- Currently in the final year of a Master's degree or engineering degree, with a solid foundation in statistics, machine learning, and programming
- Experience with Python and tabular data
- Interest in multidisciplinary projects at the intersection of statistics, regulation, and economics
- Strong analytical and problem-solving skills, with an interest in developing fair and responsible AI models
- Curiosity and desire to invest in learning the fundamentals of causal inference
- Fluent in English
Why Join Us?
- Work in a dynamic start-up environment that encourages curiosity and excellence
- Receive continuous training and development through a comprehensive learning experience that combines technical knowledge, soft skills, and practical experience
- Be part of a welcoming and close-knit community that values fun and creativity
- Imagine unexpected solutions and step out of your comfort zone to foster innovation
- Enjoy access to a range of benefits, including a private gym, art exhibitions, video games, concerts, and CSR challenges
Our Recruitment Process
- Technical test on HackerRank
- HR interview with a Talent Acquisition specialist
- Case study with a Consultant
- Final interview with a Senior Consultant
We look forward to meeting you and learning more about your passion for data science. Apply now and let's explore the exciting world of AI together!
References:
[1] Section 2 - Article 8. Requirements for high-risk AI systems: Compliance with the requirements. Of Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 laying down harmonised rules on artificial intelligence and amending Regulations (EC) No 300/2008, (EU) No 167/2013, (EU) No 168/2013, (EU) 2018/858, (EU) 2018/1139 and (EU) 2019/2144 and Directives 2014/90/EU, (EU) 2016/797 and (EU) 2020/1828 (Artificial Intelligence Act) https://eur-lex.europa.eu/eli/reg/2024/1689/oj
[2] Solon Barocas and Andrew D. Selbst. Big data’s disparate impact. California Law Review, 104-671. 2016
[3] Solon Barocas, Moritz Hardt, and Arvind Narayanan. Fairness and Machine Learning: Limitations and Opportunities. MIT Press. 2023
[4] Michael Moor. Causation in the Law. The Stanford Encyclopedia of Philosophy. 2019
[5] Joseph Y. Halpern. Actual Causality. MIT Press. 2016
[6] Drago Plečko and Elias Bareinboim. Causal Fairness Analysis. Foundations and Trends in Machine Learning: Vol. 17, No. 3, pp 1–238. 2024
[7] Victor Chernozhukov, Denis Chetverikov, Mert Demirer, Esther Duflo, Christian Hansen, Whitney Newey and James Robins. Double/Debiased Machine Learning for Treatment and Causal Parameters. ArXiv preprint. 2024. https://arxiv.org/abs/1608.00060
[8] Helmut Farbmacher, Martin Huber, Lukáš Lafférs, Henrika Langen and Martin Spindler. Causal mediation analysis with double machine learning. ArXiv preprint. 2021. https://arxiv.org/abs/2002.12710
[9] Philipp Bach, Victor Chernozhukov, Malte S. Kurz and Martin Spindler. DoubleML - An Object-Oriented Implementation of Double Machine Learning in Python, Journal of Machine Learning Research, 23(53): 1-6, 2022 https://www.jmlr.org/papers/v23/21-0862.html
[10] Maresa Schröder, Dennis Frauen, and Stefan Feuerriegel. Causal fairness under unobserved confounding: a neural sensitivity framework. 12th International Conference on Learning Representations, 2023.
[11] Julia Angwin, Jeff Larson, Surya Mattu, and Lauren Kirchner. Machine Bias: There’s software used across the country to predict future criminals. And it’s biased against blacks. ProPublica https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing. 2016
As an employer, Ekimetrics offers equal opportunities to all applicants regardless of gender, ethnicity, religion, sexual orientation, social status, disability, or age. Ekimetrics is committed to fostering an inclusive work environment that reflects diversity in its teams.