Anthoni Schiochet

Within this scale-up offering its clients a Data Catalog SaaS solution, in an R&D department of 20 people, I served for 2 years as Team Leader for a team of 6 developers. This role led me to:

- provide middle management for my team members

- participate in the development of the R&D's overall processes

- solve the organizational, operational, and technical challenges posed by the company's growth and product strategy changes

- propose technical solutions through POCs, design, architecture, and process working groups

In addition to my Team Lead responsibilities, I also contributed as a Software Engineer to developing product features, building the continuous deployment system, and improving the product's deployment infrastructure.

I am working on a Machine Learning central component used by Adevinta's marketplaces like Leboncoin, to provide them with automatic image classification and moderation features (detection of nsfw, text in image, image tagging...). The models are served by Tensorflow, and exposed to the customer through an API layer written in Java.

My contributions in this project are various, from AWS infrastructure, to Java backend development, data engineering, CI/CD automation, monitoring, architecture & performance optimizations.

- migration of the infrastructure from one AWS account to another one, provisionning of VPC , subnets, ACL, security groups and foundamental resources

- implementation of the CI / CD pipelines using Spinnaker, Travis, Terraform, Artifactory, Docker

- obervability improvements, using Datadog dashboards based on backend custom metrics, automatic alerting reported on Slack, log gathering in Sumologic

- stress & load testing, fine-tuning, cost optimization

- redesign of the system architecture, to improve runtime performances & to fix scalibilty issues

- deep troubleshooting of Kafka related issues in the central data pipeline

- close collaboration with backend engineers & data-scientists to solve challenges of deploying new models

- contributions to the creation of an global Machine Learning platform under development (ArgoCD, Kubeflow...)

- publicly presented the result of my work to a large audience, in a english-speaking environment

- got acquainted with OKR methodology

Technical environment:

Java, Python, bash, linux, AWS, Datadog, Sumologic, Terraform, Travis, Spinnaker, Docker, Packer, Tensorflow, Kafka, Artifactory,

➢ Médiamétrie develops software for reporting TV, Radio, and Internet audience measurement, intended for businesses (TV channels, radio stations, and websites).

Its IT department consists of about a hundred people (including 60 developers, 10 project managers/Product Owners, 10 production engineers, 10 system and network engineers), organized into Agile product teams, and working on AWS Cloud in a Devops environment.

➢ When I joined this company (2012), Agility was only applied by development teams, and not by Ops teams or business units; similarly, although a Software Factory existed, it only covered continuous integration aspects; integration tests were not automated, nor were UI tests; production deployments were infrequent and therefore risky, and deployment processes were manual; methods and tools differed between dev and ops; AWS Cloud was not used, let alone known by the production teams. Database migrations were executed manually.

➢ Initially hired for my development expertise, I worked on the development of the company's main application, which occupied 80% of my time. I held the position of Scrum Master for this project for almost 3 years.

➢ The development of this application also served as a support and pilot for implementing the IT department's ambitious evolutions as a whole: continuous integration, automated integration & UI tests, adoption of AWS Cloud, versioned database migration, continuous deployment, frequent deliveries, devops...

➢ In parallel, I was involved, directly or indirectly, in many small projects, task forces, "chapters", and other cross-functional working groups aimed at reflecting on the major Agile, Cloud, and Devops evolutions of the IT department, and I also ensured their concrete implementation.

➢ I also participated in defining the recruitment process, for example by developing a CodingGame technical questionnaire. I conducted over 50 hiring interviews and thus recruited over 10 employees, for the project team I worked on or for other development teams.

➢ Finally, I developed a strong awareness of operational issues (redundancy, scalability, backups, monitoring, network security,...) by working in close collaboration with expert AWS architects.

➢ Here are the main tasks I performed:

• Definition and implementation of continuous integration/deployment scripts and Infrastructure as Code (Chef, AWS Opsworks, Jenkins, XLRelease, Terraform, Ansible, Consul, Git)

• Devops: training Ops teams on development tools (GIT, IDEA) and methods (Agile, DoD, continuous integration)

• Devops: evangelizing Ops teams on Flyway

• Participation in the standardization of Devops tools for the entire IT department (Terraform POC)

• Development of the company's main web application (Java GWT Spring AWS Akka)

• Scrum Master for a team of 5-9 developers for 3 years, Project Management Assistance

• Evangelization/training of the development department in Behavior Driven Development (Cucumber)

• Development of business (Cucumber) and UI (Selenium) test scenarios

• Definition of the recruitment process and technical interviews,