Eliot B.

As a Senior Backend & Cloud Engineer, I designed, developed, and operated the backend architecture and cloud infrastructure for a multi-site research platform.

Key achievements:

➔ Design and evolution of a serverless AWS architecture:

GraphQL APIs with AWS AppSync

Node.js backend on AWS Lambda

MongoDB Atlas database

AWS Services (S3, CloudFront, Route53, SES, SQS, SSM...)

➔ Industrialization of infrastructure with Infrastructure as Code:

Terraform / Terragrunt

Management of dev, stage, and production environments

Reusable Terraform modules

Automated provisioning of cloud resources

➔ Implementation of an automated deployment platform:

CI/CD GitHub Actions

Self-hosted AWS runners

Conditional deployments based on code changes

Automation of frontend, backend, and infrastructure deployment

➔ Development of a multi-site provisioning system for creating new application instances from business configurations:

Automatic generation of necessary AWS resources

Frontend deployment via S3/CloudFront

Backend configuration and associated APIs

➔ Implementation of observability and operations:

Self-hosted Sentry for error tracking

Grafana for metrics and dashboards

Ackee for analytics

CloudWatch for logs and alerting

Main Stack:

AWS · Terraform · Terragrunt · GitHub Actions · Lambda · AppSync · GraphQL · MongoDB Atlas · Node.js · Vue/Nuxt

Design and development of backend services for the TRIBRating platform in an AWS cloud environment. Implementation and evolution of REST and GraphQL APIs, with a focus on performance, robustness, and security.

Industrialization of deployments via CI/CD (Jenkins) and implementation of infrastructure as code with Terraform on AWS. Contribution to logging and monitoring pipelines with the ELK stack (Elasticsearch, Logstash, Kibana).

Demanding environment (finance) with strong constraints on reliability, traceability, and scalability.

Design of the "RATP" (Reliable Asynchronous Transport Protocol) network protocol.

RATP aims to make the UDP protocol reliable while maintaining an easy implementation and ensuring the order and integrity of high-throughput data (10 Gb/s). Unlike TCP, a blocking protocol, RATP handles losses asynchronously; it does not stop packet transmission during losses.