Mateo Shija
AI & Data Specialist
A selection of AI and data projects — from RAG systems to annotation workflows — delivered for real-world use cases.
Built a Retrieval-Augmented Generation system allowing an organization to query internal knowledge bases via natural language in both English and Swahili.
Designed and executed a full multimodal annotation workflow for an English–Swahili code-switching NLP dataset, including QC framework and style guidelines.
Created structured image annotation datasets using bounding boxes and segmentation techniques to train object detection models for a computer vision project.
Built an automated AI-powered workflow using n8n to streamline data collection, classification, and routing for a small organization — reducing manual processing time significantly.
Conducted an AI readiness assessment and hands-on workshop for an NGO team in Tanzania, covering practical AI tools, prompt engineering, and implementation planning.
Performed comprehensive audio transcription and annotation — including timestamping, speaker labeling, and handling real-world Swahili–English code-switching scenarios.
Have a project in mind? Let's build it together.
Start a Project →