Front office (Global Market Division)

• Applying machine learning for financial markets

Front office (Global Market Division)

• Development of a tool for the rapid valuation of structured and vanilla products through the use of models that lie at the intersection between physics and deep learning.
• Master thesis: Physics Informed Neural Networks (PINN) for option pricing

MLOps: Automation and deployment at scale of various time-consuming operational financial processes:

• Automating RWA (Risk-Weighted Assets) calculation process
• Deploying at scale of the credit scoring model in collaboration with data scientists
• Automating reporting of several accounting documents (balance sheet, ledger, financial audit trail, syndication,…)

Building ETL pipelines to combine third party data from APIs with the datalake, with the objective of data enrichment for the creation of new digital financial product features in compliance with financial regulations

Led the development of a BI architecture for monitoring data quality and tracking marketing, accounting and risk activities to improve data observability and provide insights:

• Implementing Data Lineage of the reporting model
• Identifying data mart needs to improve data access
• Conducting user interviews to understand the organization’s data pain points to drive data quality monitoring and identify business KPIs
• Implementing business KPIs to monitor marketing, accounting legal, and risk activities in order to support business units managers in the decision-making process
• Implementing data quality KPIs to ensure data quality and data profiling for the monitoring of the reporting data model
  

Led release management and deployment in the production environment and maintaining the CI-CD pipeline

Designing and implementing secure data acquisition and integration strategies for a financial industry client:

• Configuring data referential mainly in the cloud and in Hadoop environment
• Building data pipelines to collect, transform and process data in collaboration with data scientists to meet the requirements of advanced analytics data modeling
• Automation and industrialization of reporting for the monitoring of risk management, sales, marketing and accounting activities

Academic project during the post master’s degree at Télécom Paris in connection with a company in the energy sector:

• Research: Writing of a state of the art document on NILM technologies for activity detection from the aggregated load curve of a house
• Analysing: Data investigation from the electrical load curve
• Modelling: Building unsupervised end-to-end model based on an convolutional autoencoder 1D architecture for activity detection from the household’s aggregated power signal
• Deploying: Implementing end-to-end solution based on a convolutional autoencoder for anomaly detection

Scientific research work on text generation and text summarization techniques:

• Research: Writing of a state of the art document on automatic text generation and summarization techniques
• NLP: Applying a transfer learning approach by fine-tuning state of the art models based on a “transformer” architecture
• Deploying: Integrating the fine-tuned model into an end-to-end web application for the automatic generation of french meeting reports

Designing, an expense report management module for consultants:

• Analysing: Designing application’s workflow
• Modelling: Designing application’s data model
• Implementation: Developing application’s frontend and backend
• Testing: Developing unit tests
• Methodology: Scrum methodology: Writing user-stories, maintaining backlogs, managing bugs and tests