Dimitri Beubry

✅ Achievements

• - Developed a template and automated email engine for a fundraise, optimizing the management of roadshows between startups and investors.

• - Designed an AI-powered automated claim system for an insurance company: capacity of +200 files/day, significant reduction in processing times and operational costs.

• - Created an automated strategic information sourcing application (founders, market, competitors…) with investment memo generation.

• - Implemented a monitoring and scoring tool for investment firms in France (collection of financial statements, public data analysis, LinkedIn enrichment), facilitating targeting and marketing campaign effectiveness.

As the first employee of the startup, I significantly contributed over 2 years to the development of a no-code software creation SaaS platform for businesses, enabling them to digitize themselves: CRM, Invoicing, Order Management, Production Management, Intervention Management, Expense Reports, e-learning, etc.

Team of 4 developers, 1 PO

In production, a thousand daily users + public API

100 requests/second peak

1 Front, 5 Backend Services

- Infrastructure design, selection, and sizing

- Design and production deployment of an event-driven architecture for running an automation scenario platform

- Setup and maintenance of continuous deployment CI/CD with Gitlab

- Load testing

- Provisioning with Ansible

- Containerization of APIs and deployment on Kubernetes

- Monitoring and alerting with Datadog

- Form editor with dynamic addition of different field types on the page

- Creation of a no-code CSS graphical customization system

- Creation of a spreadsheet view with dynamic calculations between tables, similar to Airtable (virtualization / infinite scroll) with field types. Creation of a public API in NestJS for creating Make & Zapier modules

- Creation of a scenario automation platform similar to Make (e.g., Integromat) with temporal, "onCreate", "onUpdate", and "When a condition is met" triggers

- Integration of Stripe Connect electronic payment and Yousign electronic signature as a marketplace

- Integration of the OpenAI API for AI-generated responses in text and image format

- Sending emails by allowing the Maker to specify their own SMTP

- Selection of libraries and technologies

- Migration of the codebase to Monorepo/TypeScript

- Unit tests, integration tests

As part of a redesign of its existing training tools initiated in 2018, Excilys launched in 2019 a complementary module for detailed monitoring and recurrent evaluation of employee skills. This module should allow proposing evaluations, visualizing results simply, and suggesting personalized training paths based on the results obtained and known typical paths within the system.

The team consists of seven developers and operates using the agile methodology with 2-week sprints.

The application runs on a React front-end, a back-end composed of 4 Spring-Boot microservices, and a NoSQL DynamoDB database. The front-end communicates with the back-end through a REST API called Gateway, which then calls the microservices.

When I joined the project, the back-end was already developed. I primarily worked on the React front-end. I started by coding the Header, which needed to adapt based on the pages and screen size (mobile and desktop), then I fixed a form validation bug, and finally, I created the page to list, search, and sort experts.

I then worked on the Redux actions that call the API to handle the waiting time for Failure or Success responses. This allows for standardized display of Loaders during component loading.