Elkhader Fatni

Within the Client Development and Innovation Division, the Crédit Agricole Group DataLab is a center of expertise dedicated to data science and its applications in the banking sector. Its role is to create innovative approaches for leveraging internal and external data, whether structured or unstructured. As part of its missions, high-value scientific topics are studied: Machine Learning, Auto-ML, Natural Language Processing, Process Mining, Time Series Mining, Deep Learning, Geomatics, etc.

Within the "Digibiz" team (10 people), jointly with internal partners (Regional Banks):

- Coordination of different stakeholders (business, data scientists, Data Analysts) and provision of solutions tailored to their needs

- Major participation in the design of the full-stack, API-driven architecture, and algorithm performance

- Design, sizing, and implementation of the Cloudera Dev and Acceptance infrastructure

- Design and then implementation of the Data architecture/flows within the cluster (data layer)

- Development of industrial AI solutions

- Implementation and technical lead of DevOps best practices

- Design and implementation of the project using ScrumBan methodology with business stakeholders

- Functional programming in Scala

- Facilitation and management of agile ceremonies

The business unit of the Finance – Accounting and Prudential Ratios domain is awaiting a solution and support based on BigData technologies and the associated concept of Datalake. The business has very high expectations for this datalake, which is seen as a solution accelerator for all future regulatory reporting to be implemented from 2017 to 2020, as well as for improving the quality of accounting, prudential, and risk data within Natixis's IS.

Within the Finance & Risk "Datalake" team (15 people):

- Development of innovative projects for data management (storage), processing, and reporting

- Design and implementation of these projects using Agile methodology with business stakeholders

- Implementation of Big Data and NoSQL technologies and frameworks such as Hadoop, Kafka, Nifi, Spark, Scala, Elastic Search, Hive, Kerberos, Kibana, Grafana…

- Development in Spark / Scala within an Agile team for the production of accounting reports up to the industrialization phase

- Functional programming in Scala

- Automation of data flows under NiFi and Control-M

- Use of the entire continuous integration chain in place at Natixis (DevOps), scheduling and automation of big data jobs

- Facilitation of agile ceremonies

As part of a project to overhaul BNPP Group's Operational Risk IS, a Big Data HDP architecture was chosen to form the data backbone of this new IS. This project involves IT development, IT support, and architecture teams in an agile approach. The project's objective is to provide the group's management and the bank's regulator with quality reporting and the ability for predictive analysis to control and reduce their operational risk provisions.

Within the Datalytics project (10 people, BAs, experts, developers):

- Implementation of project methodology from prototyping to industrialization

- Setup and support for business teams to data teams to achievement teams

- Study & installation/administration of Big Data tools (BigInsight platform then HDP)

- Integration of legacy data into the data lake (from CSV, Oracle databases, BO, ORX base…)

- Processing and data quality improvement using a 3-layer approach: raw-normalized-valuable data

- Support for reporting setup

- Evolution of classic reporting towards predictive reporting (high stakes around operational risk provisions):

• Learning with internal cases (supervised learning of incident data)

• Unsupervised learning (clustering of historical incidents)

- Presentation of results to stakeholders (dataviz) and business recommendations: interactive dataviz (D3.js)

- Industrialization of the POC: migration and management of prototype projects towards industrialization

- Coaching and support for agile project management

• Scrum master: daily meetings, sprint reviews, retrospectives…

• Tool supervision: JIRA Agile (Confluence)

- Management of security for KAFKA consumer/producer exchanges.

- Participation in the project to harmonize inter-application Json message formats

- Creation and standardization of a dedicated Java producer/consumer framework for developers

- Creation of Data science prototyping tools in Python

- Study of the data model in Jupyter