Ahcène Maouche

LLM Functional Testing: Analysis and validation of responses generated by "K" from patient cases, verifying reasoning quality (logical flow, sources, plots, cognitive structure). Implementation of the complete QA strategy: Design of the testing methodology (manual testing, exploratory testing, prompt-based testing, flow testing), structuring bugs, documentation, development of test plans, formalization of use cases, and prioritization of critical paths. End-to-end automation from scratch: Creation of an E2E testing framework with Cypress, integration into a GitHub Actions CI/CD pipeline, execution on Dockerized environments (Dev/Staging), coverage of critical scenarios including the complete user-LLM flow. Monitoring prompt response time: Implementation of performance metrics in Cypress scripts, analysis of latencies via JSON logs, generation of alerts in case of threshold breaches, ensuring model stability and responsiveness. Test-oriented prompt engineering: Creation of structured prompt batteries to simulate complex business cases, analysis of AI responses with biologists. Internal automation training: Implementation of training sessions for developers and business analysts on Cypress, sharing QA best practices applied to artificial intelligence, developing a test-first culture within the AI team.

Impact of the role:

Reliability of a critical-use medical LLM. Reduction of manual validation time by 50% for strategic paths, as well as better manual test structure. Strengthening of functional coverage: +50% of business cases covered thanks to the extension. Cross-functional skill development: Supporting product and AI teams towards the adoption of a rigorous, documented, and scalable quality approach.

QA Leadership: Led an international QA team to implement a testing strategy integrated into CI/CD pipelines, using Jenkins and Grafana to monitor performance. Integration of LangChain and Ragas to validate conversational flows of AI agents, applying Prompt Engineering and Fine Tuning techniques. QA Process Optimization: Contributed to obtaining ISO 9001 certification by implementing rigorous testing processes, reducing production deployment times by 20%. Design and deployment of CI/CD pipelines with Jenkins, integrating automated tests to ensure application stability before production deployment. API Testing: Designed and executed API tests with Postman to validate integrations between systems and ensure data exchange consistency. Use of AWS Bedrock and OpenAI to test the performance of intelligent agents and LLM models. Demo preparation and presentation: Prepared and presented functional demonstrations to stakeholders to illustrate progress, validate behaviors, and gather business feedback. Use of Python and Shell: Automated backend tests and analyzed logs to identify critical anomalies, using Python and Shell scripts to optimize processes. User story writing: Participated in writing user stories in collaboration with the PO. Continuous Integration and Monitoring: Configured CI/CD pipelines with Jenkins and set up Grafana dashboards to monitor application performance, in collaboration with the teams. Deliverable validation: Responsible for pre-production validation, overseeing log analysis, quality indicators, and critical tests to ensure system stability.

Automated Testing: Developed automated tests with Cypress and Python to ensure application quality, integrating these tests into CI/CD pipelines via Jenkins. This automation reduced testing times by 40% and increased test coverage by 25%. Close collaboration with DevOps teams to integrate automated tests into pipelines. Automation and CI/CD: Automated tests via Jenkins and monitored performance with Grafana. Centralized version management with Git. API Testing with Postman: Designed and executed API tests with Postman to validate integrations between systems and ensure data exchange consistency. Performance Monitoring: Used Grafana and AWS CloudWatch to monitor performance. Native mobile testing (iOS/Android) on real devices and emulators via BrowserStack, with validation of CarPlay compatibility (embedded interfaces). Requirements and Defect Management: Responsible for writing and managing test requirements and tracking defects via JIRA XRAY, in collaboration with business and technical teams. Performance Testing with JMeter: Designed and executed load and performance tests with JMeter to evaluate application scalability and robustness, simulating complex user scenarios. Use of Python and SQL: Automated backend tasks and analyzed application performance to improve stability, using SQL queries to manipulate and analyze complex datasets. Infrastructure Management with Docker, Terraform, and Kubernetes: Configured and managed test and production environments with Docker for containerization, Terraform for automatic cloud resource provisioning, and Kubernetes for container orchestration. These tools enabled environment standardization and reduced deployment times by 25%.