Zakaria MENNIOUI

- Contributed to flight schedule optimization by maximizing revenue (connection revenue and local revenue) and minimizing costs (missed bags, operational costs, etc.) while adhering to multiple constraints (minimum duration between flights, TTMs, number of gates, interventions, etc.). - Implemented heuristics to reduce time complexity. - Designed the architecture for end-to-end testing of the flight schedule optimization and automated the tests using GitHub Actions.

After completing an academic project in collaboration with VPDive, which involved detecting and classifying fish from images taken by divers, I was offered a mission to migrate the code to AWS and create an EC2 instance with optimal configuration. • I opted for a g4dn.xlarge instance, which significantly reduced the training time while maintaining reasonable costs and proceeded with the code migration. • I developed an inference script that, given an input directory of fish images, draws bounding boxes around each fish and performs their classification.

In the Big Data Lab team, under the Strategy and Transformation Department, I developed a machine learning model aimed at predicting the success of a bank representation. A bank representation corresponds to the subsequent attempt of automatic debit after a rejection. Objective: Target bank representations more precisely to maximize debt recovery and reduce the unsubscription rate. • Data enrichment with information from invoices, offers, automatic debits, etc. • Data visualization and statistical analysis of the impact of different variables on the success of representations. • Development of a machine learning model to predict the probability of success of an automatic representation. • An evaluation conducted on two control groups (representations with and without the model) revealed that the representations selected by the model performed better in terms of success rate and the amount recovered.

Within Deloitte's FSI.AI team, specializing in the development of web applications for Data Science projects, I contributed to the enhancement of their Enterprise Resource Planning (ERP). • Creation of API routes for adding/removing/modifying consultants associated with a mission. • Implementation of unit tests and end-to-end tests. • Enhancement of code quality by adding logs, custom exceptions, etc. • Initiation of project documentation

Within the production team at Redslim, I contributed to the extraction, aggregation, and transformation of Consumer Packaged Goods (CPGs). • Implementation of massive data transformations and quality checks (QC) using Spark. • Deployment of new functionalities into production using Azure Pipelines and Azure Data Factory. • Implementation of unit tests and enhancement of code quality.

Within the technical team at Minautor, I contributed to the development of the website's backend as well as the creation of some internal tools. • Implementation of an API using FastAPI enabling the search for used car engines by brand and model. • Data collection on car engines from the internet to enrich the internal database. • Development of chatbots for Minautor on Telegram and WhatsApp

Skills: Git · Django · Python (Programming Language) · Communication