Quantitative AI Portfolio Engineer (Fixed Income) H/F

Amundi; Paris; Hybride; EUR 90 000 - 130 000; Plein temps

Résumé du poste

Amundi recherche un ingénieur Portfolio IA Quantitatif pour concevoir, valider et déployer des modèles d'investissement basés sur l'apprentissage automatique dans le Lab Fixed Income Investment.

Vous travaillez sur le cycle complet: ingénierie des données, construction de features, back-testing et industrialisation, avec une attention à la robustesse et à l'explicabilité des signaux.

Qualifications

• Diplôme avancé en finance quantitative, ingénierie financière, mathématiques appliquées, informatique, data science ou équivalent.

Responsabilités

• Concevoir et développer des modèles ML pour générer des signaux d'investissement sur les taux et le crédit.
• Réaliser des feature engineering sur macroéconomie, microstructure, flux, ESG et sentiment.
• Construire des bibliothèques réutilisables de features et modèles et des outils de validation.
• Mettre en place des back-tests robustes (look-ahead, coûts de transaction, slippage).
• Collaborer avec Amundi Institute pour transformer les avancées académiques en modèles opérationnels.
• Surveiller la dérive des données et des modèles, documenter les méthodologies et la gouvernance.

Connaissances

Python

Machine learning

NLP

Deep learning

Transformer models

PyTorch

XGBoost

SQL

C++/Java

Cloud platforms

Formation

Advanced degree in quantitative fields

Outils

Airflow

Spark

Kafka

AWS

GCP

Azure

PyTorch

TensorFlow

spaCy

Transformers

pyspark

Description du poste

CDI

06/07/2026

Non

Cadre

Missions

• Design and develop quantitative approaches and machine learning algorithms to generate investment signals on interest rates (curve dynamics, term premia, vol, macro linkages), and credit markets (spread dynamics, momentum, carry, liquidity, regime effects).
• Perform advanced feature engineering (macro, market microstructure, flows, liquidity, ESG, sentiment).
• Develop supervised and unsupervised learning models:
  • tree-based models, ensemble methods,
  • deep learning architectures (LSTM, Transformers when relevant).
• Integrate alternative and unstructured data (news feeds, central bank communications, broker research, transcripts, regulatory publications).
• Build & maintain a reusable internal library of features, models, preprocessing pipelines, validation tools.
• Collaborate with the research ecosystem (e.g. Amundi Institute) to translate academic innovations into operational investment models.
• Implement and oversee robust back-testing frameworks addressing biases (look-ahead, survivorship), transaction costs, liquidity constraints and slippage.
• Perform deep robustness analysis via stress tests, walk-forward analysis, bootstrap methods and stability checks across market regimes, including crisis periods.
• Measure risk-adjusted performance of strategies and evaluate sensitivity to macro and market factors.
• Define clear model acceptance criteria, rejection thresholds and degradation metrics.

Engineering & Productionisation

• Own the full model lifecycle: specification, prototyping, validation, industrialization, monitoring and maintenance.
• Implement monitoring and alerting for data drift, model drift and performance decay, and define rollback procedures.
• Document methodologies, assumptions, validation metrics and production procedures.
• Prepare reports and presentations for senior management, investment committees and client-facing teams.
• Communicate model rationale, risks, limitations, and governance aspects clearly to non-technical stakeholders.

Actively contribute to model governance, internal audits, regulatory reviews when applicable.

Within the Fixed Income Investment Lab, the Quantitative AI Portfolio Engineer is responsible for the research, development, validation and deployment of machine learning–driven investment models for fixed income portfolios.

The role focuses on end-to-end signal generation on rates and credit markets, from data engineering and feature construction to back-testing, industrialization and continuous monitoring.

The position sits at the intersection of quantitative research, applied AI and portfolio management, with the objective of transforming data into robust, explainable and actionable investment signals integrated into the daily decision-making process of Portfolio Managers.

Paris

Advanced degree in quantitative finance, financial engineering, applied mathematics, computer science, data science or equivalent.

3 - 5 ans

Significant experience (1+ to 6-7 years) in quantitative research, data science applied to financial markets, or systematic strategy development.

• Strong knowledge of modern machine learning techniques (neural networks, XGBoost, sequence models such as RNN/LSTM/Transformer, ensemble methods).
• Practical experience in NLP (Transformers (BERT), embeddings, fine-tuning, Clustering, Classification and sentiment analysis) applied to financial text.
• Solid understanding of fixed income markets (yield curve structure, credit spreads, interest-rate derivatives) and portfolio constraints.
• Strong programming skills in Python (Pandas, scikit-learn, hugging face, spaCy, sentence-transformers, PyTorch/TensorFlow, pyspark), SQL. C++/Java knowledge is a plus.
• Familiarity with data engineering tools (Airflow, Spark, Kafka) and cloud platforms (AWS/GCP/Azure) is advantageous.
• Scientific mindset, intellectual curiosity and experimental rigor.
• Ability to synthesize and explain technical results to non-technical audiences.
• Autonomy, initiative and strong collaboration in cross-functional teams.
• Production-oriented mindset with focus on reproducibility and operational robustness.

Langues

Fluent in French and professional proficiency in English.