Tanguy Epifanic

Contexte : Mission de modernisation d'un applicatif métier critique pour le compte de l'OCIRP, organisme d'assurance de personnes et de protection sociale. Le projet était bloqué sur Angular 7 par Visual Guard 2019, qui interdisait toute montée de version Objectif : Refonte du système d'authentification et des habilitations, migration de la stack vers les dernières versions, sans interrompre la TMA Équipe : 5 personnes : 3 développeurs sur un scopé donné (TMA, migration), 1 CTO intervenant ponctuellement, 1 directeur de projet. Principalement sur la migration du système d'authentification et de .NET / Angular Mon rôle : Migration du système d'authentification de Visual Guard vers Microsoft Entra ID, via un outil console .NET développé pour l'occasion : extraction du référentiel, réconciliation des anciens rôles et groupes avec le nouveau modèle, injection automatisée en base Accompagnement du client sur la bascule de ses utilisateurs, MFA activé sur les groupes le nécessitant Création d'un back-office web en remplacement de la console Visual Guard : rôles, permissions, périmètres, délégations et journalisation des connexions Migration du front de Angular 7 à Angular 22, chiffrée par mes soins Participation à la migration du back d'une architecture en micro-services de .NET 2.2 à .NET 10 Configuration de la pipeline GoCD et des environnements internes (VM dev / int / release) Entretiens techniques de recrutement et chiffrage d'avant-vente Résultats : 800+ utilisateurs basculés sans interruption de service, livré en 6 mois recette client comprise Chargements réduits de 50%, compilation ~20% plus rapide, déploiement ramené de 45 à 25-30 minutes 18 rôles, 315 permissions et 3 assureurs détenant de 5 à 100 délégations, gérés en autonomie par le client Création d'un profil utilisateur passée d'un ticket de plusieurs jours à quelques minutes Technos & outils : C#/.NET | Angular | TypeScript | Microsoft Entra ID | MSAL | Visual Guard | T-SQL | GoCD | Micro-services

Context: Modernization mission for a critical application in the insurance sector, suffering from extreme loading times (up to 57 seconds). Objective: Optimize back-end and front-end performance, clarify API flows, and clean up the database. Team: 5 people: 1 tech lead, 1 principal engineer (.NET), 1 junior developer, 2 interns, 1 PM. Close collaboration with the tech lead, mentoring junior profiles, and strong support on .NET topics (team mainly from the Java ecosystem). My role: Complete rewrite of API calls via a Back For Front (Ocelot + Proxy) - 250+ REST APIs rewritten Creation of batch API REST endpoints rather than individual ones Creation of batch API aggregations Creation of optimized pagination with EF Core Front-end refactoring to integrate new optimized call patterns Data model cleanup (database optimization) Mentoring junior profiles/interns Improvement of internal processes (code review organization, ticket management) Team introduction to generative AI Results: Results expected by September Strengthened client confidence and willingness to continue collaboration Technologies & tools: C#/.NET | Entity Framework Core | VueJS | Azure DevOps | AWS (Amazon Web Services) | Postgre SQL | SQL Server | GitHub | Ocelot | CQRS | MediatR | Postman

Contexte : Mission de création d'application moderne from scratch pour le compte d'Edmond de Rotschild Objectif : Création d'un outil d'analyse de prospectus financier via Azure Databricks, avec validation manuelle utilisateur et export en Excel Équipe : Seul développeur du projet Mon rôle : Création d’une interface moderne et ergonomique avec Angular 20 et Angular Materials Connexion avec un monolithe .NET en clean architecture avec qualité du code Analyse de prospectus via Azure Databricks Résultats : Projet livré en janvier Temps de traitement des prospectus financier divisé par 2 Technos & outils : C#/.NET | Angular | Azure DevOps | SQL Server | Clean architecture

Context: Reinforcement on a strategic project to modernize a tax declaration tool for businesses. Main objective: Reducing technical debt, improving code maintainability, and transitioning to a more efficient modern stack (Razor -> React). Team: Team of 9 people on the IFU (Unique Tax Identifier) module: 1 tech lead, 5 developers (senior, intermediate, apprentice), 1 PM, 1 Scrum Master, and an external designer. Close collaboration with the PM to ensure functional adequacy of deliverables. My role: Legacy code refactoring (dead code analysis, strangler fig pattern implementation) Complete migration from Razor to React Critical SQL query optimization Sonar compliance (comprehensive correction of critical and major alerts) Contributing to the implementation of generative AI to simplify tax declaration input Implementation of an effective code review strategy Facilitating agile ceremonies Results: -40% cognitive complexity Test coverage increased to 80% Loading time cut in half Clear documentation and shared best practices to ensure technical continuity Technologies & tools: C#/.NET | React | SQL Server | Azure DevOps | SonarQube | SonarLint | Moq | Jest | Postman | Redis | Dapper | TypeScript | JavaScript | Kendo UI | REST API

Context: Mission focused on improving acquisition and conversion for Le Figaro's job platform Objective: Increase SEO traffic and modernize user experience through Vue 2 → Vue 3 migration, while ensuring the team's technological transition and staying alert on the site's status during traffic spikes. Team: 1 tech lead, 3 intermediate developers initially, then final pair with the tech lead. Direct collaboration with product teams to prioritize business needs and guide deliverables. My role: Complete interface migration: Vue 2 → Vue 3, Bootstrap → Vuetify, adoption of the Composition API UI component redesign to improve maintainability and performance SEO strategy: automatic sitemap generation, creation of optimized landing pages, integration of a keyword ranking algorithm, implementation of custom Elasticsearch aggregations CDN deployment on GCP to optimize media loading Compliance with RGAA & WCAG accessibility standards Deployment of REST APIs adapted to business requirements with Swagger documentation Results: +35% organic traffic -60% image loading time Test coverage increased to 80% Direct improvement in conversion on job pages Technologies & tools: Go | Vue 3 | Vuetify | Nuxt | JavaScript | TypeScript | Elasticsearch | GCP | GitHub | Docker | GraphQL | Kubernetes | Jest | Redis | WSL | Agile | Datadog | REST API | Swagger

Context: Reinforcement mission on a product in production, dedicated to advertising report management and billing. Objective: Improve performance, strengthen code quality, and support the server transition to Google Cloud. Team: 7 people: 1 tech lead, 2 senior developers, 1 apprentice, 1 QA, 1 PM. Direct collaboration with the PM and inter-team coordination in the context of cloud migration. My role: Development of reusable UI components via Storybook Creation of a dynamic "ads.txt" page for managing whitelisted advertisers Virtual scroll integration on advertising reports → loading time divided by 5 (from 5 sec to < 1 sec) Resolution of critical bugs and participation in adding new features SQL query optimization and contribution to CI/CD pipeline improvements Support for load testing in the context of GCP migration Continuous Angular version migration Results: -80% on advertising report display time Test coverage increased from 50% to 80% (minimum acceptable) Shareable UI components deployed to accelerate future development Facilitated cloud integration and more stable CI/CD Technologies & tools: C#/.NET | Entity Framework Core | Angular | GCP | SQL Server | PostgreSQL | GitLab | TypeScript | JavaScript | Datadog | Agile | ArgoCD | Storybook | REST API

Développement d’un outil web permettant la création simplifiée de pages web en no-code. Participation régulière à des évènements.

Answering tickets from the platform users and verification of identity documents of drivers who registered on the platform.

Creation of a monthly statistics web page. Representation of these statistics in the form of tables and graphs.