Understand the specific challenges associated with the task of fixing code errors.
Propose novel optimization approaches to improve the effectiveness of generative models in correcting code errors, using advanced machine learning techniques.
Implement and experiment with the proposed new approaches using existing code data and generative modeling tools such as Llama-3, Mistral, etc.
Evaluate proposed new approaches in terms of effectiveness and performance, using appropriate metrics.
Integrate the models into our production DevOps chains.
Communicate and present results in an innovative way for other teams.
Requirements & Skills:
Student in a university or engineering school course, looking for an apprenticeship contract for a minimum period of 12 months.
Specialty(ies) / Desired Option(s): Data Science, Machine/Deep Learning, Big Data, Statistics
Understanding transformer-based models
Mastery of OO programming: Python
Mastery of libraries dedicated to the development of Machine Learning and Deep Learning algorithms (scikit-learn, tensorflow, pytorch, transformers, etc.)
Mastery of Git and Gitlab tools
Good level of written and spoken English
Capacity for analysis, synthesis, and formalization
Willingness to be involved and take initiative, and autonomy