Worked on security-focused enhancements for Docker-based deployments in the zama-ai/concrete-ml and zama-ai/concrete repositories. Addressed vulnerabilities by updating Dockerfiles, introducing non-root user execution, and standardizing base images to strengthen security posture. Improved CI/CD reliability by configuring Trivy to ignore non-critical HEALTHCHECK findings, reducing unnecessary privilege requirements and supporting more stable builds. Leveraged skills in Docker, Linux, and CI/CD pipelines, with changes implemented using Dockerfile and Shell scripting. These efforts reduced the attack surface and aligned security practices across both repositories, resulting in more secure, consistent, and efficient deployment workflows for machine learning and cryptography-related projects.