Alexandre Wilmet

• research, collection, and formatting of time series and indicators

• creation of real-time update methodologies,

• design of sophisticated learning techniques, training sets, and metrics,

• development and deployment of deep learning models using RNNs, MLPs, and embeddings, aimed at predicting market conditions at key moments,

• implementation of procedures for effective hyperparameter testing and training analysis, leading models to performance significantly superior to baselines such as random.

• analysis and formatting of relevant data from mobile game users and advertising space purchases/sales,

• creation, development, and deployment of probabilistic algorithms for the automation of advertising space purchases/sales using this data,

• collaboration with technical teams for the creation and improvement of these algorithms, leading to a significant increase in performance indicators, as well as a reduction in maintenance costs through automation,

• research, development, and deployment of deep learning models, using MLPs and embeddings, to predict the probability of a user installing, achieving results equivalent to those of a competing probabilistic algorithm.

• data manipulation in an industrial setting,

• implementation of a probabilistic model,

• recognition of faces from television broadcasts with images coupled with broadcast metadata.