Armand Dadoun

• Project Competitor Footprint:

- Led a team of 10-15 developers to build a data pipeline to identify the presence of other Internet Service Provider (ISP) through many different datasource including their own public API

- Engineered a presence absence background machine learning model to predict competitor insights at specific location

- Generated increase of Monthly Recurrent Charges of more than 700K$ annually

• Project Next Best Collection Action:

- Worked as the primary Senior data scientist in developing a neural network model to predict optimal collection actions for the payment collection team

- Successfully generated over $600,000 annually in benefits by enhancing collection efficiency and effectiveness

- Leveraged data analytics and machine learning to improve decision-making and streamline operations

• Project Research and Development Project Subvention (R&D)

- Developed an AI-driven predictive model leveraging a LightGBM (LG Boost) framework to identify projects with the highest likelihood of securing government subventions. This model was integral to optimizing our Research and Development (R&D) funding strategy, enabling the identification of high-potential projects with precision. By applying the model, our team successfully secured more than 40% in government subventions for eligible R&D initiatives in the past year, significantly boosting innovation funding and maximizing resource allocation efficiency.

• Team lead and data scientist leveraging machine learning algorithms to develop a model that predicts the next best fibre builds (NBFB). This, through a non-linear regression based on several indicators (distance to Bell Tower, speed of service, etc.) This classification XGBoost model is now being utilized by all the implementation managers and helped the company generate 2M$ extra yearly revenue compared to the control group.
• Building effective ETL pipeline to real-time dashboard tools to support the operational teams run the business through cloud tools (GCP, AWS).
• Reporting automation and workflow enhanced efficiency has allowed to save 300K$ yearly of labor cost.

• Use of the stochastic & actuarial calculation methods to determine the optimal number of agents needed in a call center
• Great knowledge of the WFM process
• Starting with the call placement strategy, then the requirements forecast leading to the creation of schedules in the very well known WFM tool, IEX
• Use of the econometric theory & multiple python libraries to generate call volume forecast by interval of 15 minutes
• This represents a primary role within the organization, since the budget, financial plan and customer experience rely on this metric
• Able to automate reporting workflow and reduce yearly labor cost by 500K $ (~5 Resources)

• Predict financial performance with high-frequency data using a new measure of liquidity
• Wrote and presented a 150 pages that is successfully publicly published
• Thorough knowledge of MATLAB
• The supervisor is Alexandre Roch, tenured UQAM professor