Anass AKRIM

• Develop and structure the AI strategy for offers, products, and technologies.

• Work closely with data architects, big data specialists, and cloud architects to create and deliver products, projects, and services that meet the needs of our internal clients.

• Provided internal training, technology watch and conducted proof-of-concepts.

Benchmarking and integrating multimodal features (image & document processing) into an internal API platform (chatbot).

Co-developed an innovative Generative Design solution with Veolia and AWS, optimizing facility layouts using advanced algorithms and deploying the solution on AWS with GitLab CI/CD integration:

• Defined client requirements and scoped the optimization problem.
• Translated the optimization challenge into a mathematical model for implementation.
• Developed and implemented the optimization tool and algorithm in Python.
• Collaborated with Veolia and AWS teams to develop and deploy a chatbot powered by Agentic LLM via AWS Bedrock (Anthropic Claude 3), integrating the optimization algorithm to enhance its functionality.

Co-developed an innovative Generative Design solution with Veolia and AWS, optimizing facility layouts using advanced algorithms and deploying the solution on AWS with GitLab CI/CD integration:

• Defined client requirements and scoped the optimization problem.
• Translated the optimization challenge into a mathematical model for implementation.
• Developed and implemented the optimization tool and algorithm in Python.
• Collaborated with Veolia and AWS teams to develop and deploy a chatbot powered by Agentic LLM via AWS Bedrock (Anthropic Claude 3), integrating the optimization algorithm to enhance its functionality.

I joined the Square Research Center, the applied R&D branch of Square Management, in which I lead and participate in several projects aimed at developing innovative AI tools to address organizational challenges, in collaboration with researchers, consultants and various business partners for concrete implementations.

Adjunct lecturer at French engineering schools (ISAE-Supaero & INSA Toulouse @Mathematical Modeling & AI Department) in applied mathematics and artificial intelligence domains.

Courses I have taught include:

● Statistics and Probability
● Machine Learning
● Applied Mathematics and Optimization
● Signal Processing
● Linear programming
● Programming in Python, R and Matlab

Thesis carried out at the Institut Clément Ader and at ISAE-SUPAERO, a world leader in aerospace engineering higher education. My thesis focused on the development of a cutting-edge AI methodology for predicting the remaining useful life of aircraft components, with the aim of improving maintenance efficiency and safety.

● Physics-based modeling: Development of an innovative framework for large-scale synthetic data generation, facilitating comparative analysis in Machine Learning for predictive maintenance prognostics.
● Evaluated and benchmarked the latest Machine/Deep Learning techniques for Time Series Forecasting (i.e. prognostics in Predictive Maintenance - remaining useful life prediction).
● Proof-of-Concept : development of an innovative ML methodology allowing AI to learn from available data without external annotations - enhancing the results of prognostics in predictive maintenance despite the limitation of labeled data.
● Supervised and mentored a team of 3 students/interns, guiding them through Python programming and Machine/Deep Learning projects focused on predictive maintenance, emphasizing practical application and collaborative problem-solving.

Artificial Intelligence in Financial Forecasting:

● Developed expert-based techniques and Machine/Deep Learning models to forecast Forex market price fluctuations using Python.
● Coaching in Python programming for various banking professionals.

Machine Learning in Fraud Detection:

● Evaluated and benchmarked the most advanced unsupervised Machine Learning techniques for anomaly detection using Python.
● Developed a robust deep learning method for data drift detection in an unsupervised manner (Python).

Research on Dynamic Data Driven Forecasting for Financial Time Series Data using Machine Learning and Deep Learning techniques :

● Filtered noisy financial time series using Singular Spectrum Analysis (Python)
● Explored the most cutting edge Machine and Deep Learning algorithms for forecasting of stock prices (on Python), using also Sentiment Analysis (NLP)
● Developed automatic trading algorithm using neural networks (deep reinforcement learning, equity curve)

'Mines Etudes et Projets' is a local non-profit organization entirely executed by students of "L'Ecole des Mines de Saint-Etienne". Related to the field of studies (Computer Science, Energy, Finance...), our Junior Entreprise works as a small company and offers consulting services to the market. By doing this, the students of the School can do professional project work and add practical experience to their theoretical skills and bridge the gap between School and the Business world.

Missions :
● Managing a team of 13 people
● Developing partnerships
● Managing projects.

- Period of professional immersion in the banking sector.