Taoufik El Khaouja

Mindee provides API solutions to extract key information in scanned and pdf documents.

• ● Designed and implemented pipelines for data preprocessing

• ● Trained and deployed state-of-the-art deep learning models (Computer vision, NLP) for automatic extraction of key data from documents. Building reproducible and automated training pipelines using ClearML

• ● Finetuned multimodal LLMs

• ● Implemented AI agents and RAG with PydanticAI.

• ● Implemented post-processing algorithms to refine combined ML model outputs.

• ● Created CI/CD workflows using GitHub Actions for automated testing and quality checks.

• ● Delivered technical presentations and reports to stakeholders explaining performance, limitations and ROI

• ● Collaborate with the product team to define technical architecture decisions . Skills: LLMs, AI Agents, MLOps, CI/CD, AWS, Solution Design

• ● Training interpretable ML models (spline, general additive models) on real estate data for price estimation

• ● Built an interactive UI for real estate data visualization and analysis using Angular Skills: Python, Git, GCP, Kubernetes, Angular

• ● Studied asymptotic behavior of random walks on random graphs, deriving an upper bound for the heat kernel