Vibe Coder / AI Agent Pilot

Project overview

As part of an internal program focused on AI innovation and agentic development methods, a cross-functional initiative is launching a broad wave of technical reliability and compliance.

The challenge involves the quality reassessment of about twenty internal applications already in production, hosted on the group's AWS cloud and deployed via the internal software forge (GitLab CI / Software Factory).

The operating method is based on a paradigm shift: you don't code manually alone. You act as a true technical conductor, piloting several Claude Code agents in parallel, defining their context, rigorously reviewing deliverables, validating tests, and ensuring production deployment without service interruption.

Your missions:

Expected deliverables:

Organization & Mission Pace:

Profile sought

• Flash Audit & Inventory
• Map recurring anomalies, technical debt, and lack of test coverage.
• Identify obsolete or vulnerable dependencies.
• Monitor supervision alerts, backup status, infrastructure cost optimization, and security gaps.
• Multi-Agent Orchestration (Claude Code)
• Break down technical needs into independent and parallelizable tasks.
• Prepare and inject the appropriate execution context for each agent.
• Meticulously audit, correct, and validate the generated code and scripts.
• Application Quality & Non-Regression
• Address bugs at the root (no superficial fixes).
• Write and integrate automated non-regression tests.
• Secure critical user journeys and ensure reliable error case management.
• DevOps Quality & Cloud Security (AWS)
• Review and standardize environment configuration.
• Configure log centralization and alert routing.
• Automate and test backups with documented restoration procedures.
• Reduce the attack surface according to the principle of least privilege (role/permission management) and eliminate superfluous infrastructure expenses.
• Industrial Delivery & Capitalization
• Deploy via GitLab pipelines and the company's software factory (all systems go).
• Validate production versions with tested and operational rollback procedures.
• Write reusable instructions (system prompts, contextual guidelines) and technical procedures to sustain the approach after the mission.
• Quality dashboard before/after per audited application.
• Fixes and test suites successfully deployed in production.
• Sanitized server/AWS environments with proof elements (logs, alerts, restoration tests).
• Report of residual gaps and recommended action plan.
• Documented agentic toolkit (sample prompts, orchestration guides) for the internal team.
• Days 1 and 2: Familiarization with the pre-configured workstation, initial audit, and roadmap validation with the technical team.
• End of Week 1: Launch of the first agents, validation of the first tasks, and first code/test deliveries to production.
• Week 2: Intensive parallelization of several agents on code and servers, consolidation, production deployment, and handover.
• Rituals: Short daily synchronization meeting, weekly demonstration, and quick decision-making process during the day.
• Vibe Coding Culture & AI Agent Mastery: Intensive daily practice with Claude Code (or similar agents), high rigor in context breakdown, and ability to manage multiple simultaneous sessions.
• Solid Software & DevOps/Cloud Foundation: Excellent understanding of modern application architectures, the AWS ecosystem, CI/CD pipelines (GitLab), and production operability requirements.
• Demanding & Review Culture: Ability to quickly and accurately audit a set of changes, detect side effects or technical hallucinations.
• Experience in Corporate Environments: Strict adherence to delivery processes, security policies, and infrastructure constraints of a large group.

Project details

• Start: ASAP
• Duration: 15 jours
• Remote: hybrid
• Location: Île-de-France, France, FR
• Languages: French