AI Expert – DevSecOps

Keoni; Paris; Hybrid; EUR 90,000 - 130,000; Full time

Job summary

Keoni in Paris is seeking an AI Expert – DevSecOps to help integrate agentic AI into the CI/CD pipeline and tooling, with a strong emphasis on GitLab. You will prototype, deploy AI-enabled components, and collaborate with cross-functional teams to translate business needs into technical solutions.

The ideal candidate has 5+ years in AI development and DevSecOps, solid scripting in Python, and experience with Docker/Kubernetes.

Qualifications

• 5+ years in AI, DevSecOps, or ML-enabled development.
• Experience integrating AI into CI/CD workflows.
• Ability to communicate with non-technical stakeholders.

Responsibilities

• Create spikes and POCs around agentic AI: LLM, agents, RAGs.
• Integrate AI into the CI/CD pipeline and DevSecOps tooling (GitLab).
• Design AI enablers and reusable tools for train teams.
• Develop components: scripts, integrations, APIs.
• Acculturation and dissemination with workshops and Confluence docs.
• Support train teams in adopting AI usage and knowledge sharing.

Skills

Agentic AI

LLM & Agents

RAG & MCP

Prompts & Code Generation

AI in CI/CD

Business communication

REST APIs

Python

Scripting

Tools

GitLab

GitLab DAP

GitLab Duo

Docker

Kubernetes/OpenShift

AWS

GCP

Confluence

Jira

Job description

AI Expert – DevSecOps Context: The role requires expertise in artificial intelligence, particularly agentic AI: LLM, agents, MCP, skills, RAG, prompting, and code generation, within the team. The team explores, prototypes, and tools AI applications with a dual objective: • To leverage AI for the continuous integration pipeline and tooling, primarily GitLab; • To broadly familiarize the team with these technologies. The ideal candidate is technically skilled but also capable of understanding business challenges and communicating effectively with non-technical stakeholders.MISSIONS • Creation of spikes and POCs around agentic AI: LLM, agents, RAGs, MCPs, prompting, code generation; • Integration of AI into the CI/CD pipeline and DevSecOps tooling: GitLab, GitLab Duo and DAP in particular, MCPs on the pipeline tools ; • Design of AI enablers and tools reusable by the train teams; • Implementation and evaluation of the group's AI building blocks: Google Gemini, Vertex AI, Antigravity, Anthropic, Mistral; • Design of AI agents and assistants applied to development and DevOps; • Implementation of RAGs and use of knowledge bases; • Technological monitoring and evaluation of AI solutions relevant to the train; • Consideration of security and compliance issues in the use of AI – DevSecOps; • Development of necessary components: scripts, integrations, APIs; • Acculturation and dissemination within the train network: training, demonstrations, workshops; • Writing and popularizing associated documentation in Confluence, for technical and non-technical audiences; • Supporting train teams in adopting AI usage; • Contributing to knowledge sharing and collective skills development.Essential Technical Skills – Mandatory Requirements • Mastery of agentic AI concepts: LLM, agents, MCP, RAG, prompting; • Ability to design and integrate AI solutions: agents, RAG, model API calls; • Strong fundamentals in AI-assisted development and code generation; • Understanding of the link between AI, CI/CD, and DevSecOps tools, particularly GitLab; • Ability to understand business challenges and communicate them clearly to non-technical audiences; • Proficiency in REST APIs and AI service integration; • Knowledge of the group's AI offerings: Google Gemini, Vertex AI, Antigravity, Anthropic, Mistral; • Skills in evaluating AI solutions: relevance, robustness, and performance measurement; • Knowledge of AI assistants for development: GitLab Duo, GitLab DAP; • Awareness of security and compliance issues related to AI – DevSecOps.Other key technical skills : • Development, particularly Python, and scripting; • General IT technical knowledge: APIs, architecture, with an operational perspective; • Knowledge of AWS and GCP clouds; • Understanding of containerization and execution environments: Docker, Kubernetes/OpenShift; • Knowledge of Confluence and Jira; • Experience with agile methodologies (Scrum, SAFe) and teamwork; • Proficiency in Microsoft Office Suite, Outlook, and Teams; • Ability to explain complex concepts clearly and facilitate workshops; • Understanding of how a DevSecOps platform works.Technological Environment: A good understanding of the following ecosystem is expected, without necessarily having in-depth expertise in each component: • LLM, agents, MCP, RAG, prompting, code generation; • Google Gemini, Vertex AI, Antigravity, Anthropic, Mistral; • GitLab Duo and GitLab DAP; • Python and API integration; • Docker, Kubernetes/OpenShift; • GitLab CI/CD, Jira, Confluence; • AWS, GCP, S3NS; • Security and compliance applied to AI; • Scrum, Kanban and SAFe; • DevSecOps and SRE best practices.Desired Profile: The candidate must have a proven technical foundation in development and DevOps with 5+ years of experience. Practical experience implementing agents, RAGs, or MCPs, or contributing to AI projects, is a significant advantage. Certifications are welcome.Expected qualities: • Ability to take initiative and propose decisions and action plans; • Ability to anticipate technical impacts or associated risks and to raise alerts; • Ability to understand business challenges, analyze needs, and be proactive in suggesting uses of AI; • Knowledge of IT and telecommunications security standards and rules; • Customer-focused approach, active listening, team spirit, and good communication skills; • Strong ability to learn, simplify, and share knowledge; • Versatility, autonomy, organization, interpersonal skills, customer service orientation, adaptability, analytical and synthesis skills; • Rigor and organization; • Strong analytical and writing skills, including the ability to explain complex topics clearly to non-technical audiences; • Curiosity, openness, and a collaborative mindset.Application: CV + cover letter + copies of diplomas to be sent to [e-mail]