Khalil Ezddini

As an AI Intern at Aimigo, I work on the processing and serialization of unstructured educational data, primarily from PDF documents, using multimodal models and agents. I contribute to the evaluation and ranking of prompt–AI pairs through dedicated testing frameworks and support the integration of Aimigo's generative and expert AI capabilities into educational solutions. In parallel, I develop agentic systems and agent harnesses to orchestrate, test, and evaluate AI agent workflows.

• • Engineered and guided the development of an end-to-end AIPlatform (Next.js/Flask) to better connect startups with investors via intelligent matching and analysis.

• • Performed web scraping with Playwright and cleaned/normalized data to maintain a reliable and up-to-date source for the system.

• • Integrated VectorSearch and orchestrated AIAgents (leveraging Llama LLMs) to perform tasks such as matching, analysis, lookups, and personalized recommendations.

Developed an AI-powered customer support chatbot to automate the handling of frequent customer queries and operational tasks. Built and integrated generative AI capabilities enhanced with tool use to provide accurate, actionable responses. Contributed across the full stack using Symfony, PHP, and NestJS, while applying AI engineering practices and leveraging Gemini models.