Senior Full Stack Azure & Agentic AI Engineer

About the role

ODDO BHF Tunis is looking for a Senior Full Stack Azure & Agentic AI Engineer to design and develop enterprise‑grade applications based on Azure Cloud, Generative AI and Agentic AI architectures. This role focuses on solution design, software engineering and AI implementation within a highly regulated banking environment, collaborating with IT Operations, Security, Business and Architecture teams while respecting segregation of duties.

Key responsibilities

• Design scalable and secure Azure‑native application architectures aligned with enterprise standards.
• Define technical solutions that incorporate RAG, MCP and multi‑agent AI patterns.
• Develop backend services using Python, FastAPI and REST APIs.
• Build modern front‑ends with Angular and TypeScript.
• Create integrations with enterprise systems, APIs and data platforms.
• Ensure software quality through Test‑Driven Development, automated testing, code reviews, ADO pipelines and CI/CD best practices.
• Design and implement AI assistants, agent‑based solutions and RAG pipelines.
• Implement prompt engineering, orchestration and agent communication patterns.
• Measure AI quality using evaluation, observability and monitoring frameworks.

Required profile

• Senior Engineer or Lead Developer with strong Azure, Full‑Stack and Agentic AI expertise, able to deliver production‑ready AI solutions in a regulated banking environment.
• Demonstrates strong engineering discipline, autonomy, clear communication and effective collaboration with IT Operations, Security, Architecture and business stakeholders.

Required skills

• Python
• FastAPI
• Angular
• TypeScript
• SQL
• PostgreSQL
• REST APIs
• Clean Architecture
• Domain Driven Design
• Test Driven Development
• Automated testing
• React (optional)
• Vue.js (optional)
• Azure
• ADO pipelines
• Git workflows
• Terraform
• Openshift (or equivalent)
• LangGraph
• LangChain
• RAG
• MCP (Model Context Protocol)
• Vector databases
• LLM integration
• Evaluation frameworks