Tatyana Amugo

Ran discovery sessions with caseworkers and policy stakeholders, scoped and delivered a two-week proof of concept for an AI pipeline to automate manual document review. Built a retrieval setup with ChromaDB and Azure OpenAI, reduced processing time by 79%, and added evaluation checks to prevent confidently-wrong outputs from reaching caseworkers.

Explained complex technical products in plain language to customers with varied technical backgrounds, achieving a 96% satisfaction score. Delivered daily live product demonstrations, adjusting the pitch in real time based on customer needs.

Built automated workflows in ERPNext connected to SQL-based reporting to cut manual data entry time by 30%.

Built a Python and SQL KPI tracking model to improve marketing campaign targeting by 15% by identifying customer segments responding to specific campaigns. Presented findings directly to the marketing team in an actionable format.

Activities and societies: Coursework: LLM systems, RAG architecture, ML pipeline design, deep learning, NLP, MLOps.

Graduate Certificate in Artificial Intelligence & Machine Learning (Co-op) at Fanshawe College, covering LLM systems, RAG architecture, ML pipeline design, deep learning, NLP, and MLOps.

Activities and societies: Thesis: built a 3D CNN model to detect anomalies in real-time video, from initial design through to a working deployment.

Bachelor's Degree in Informatics & Computer Science at Strathmore University, including a thesis on using a 3D CNN for real-time video anomaly detection from design through deployment.