Applied machine learning in real systems.
MSc graduate in Information Technology specialising in AI and NLP. Built a context-aware location extraction pipeline for Finnish news articles as part of my Master's thesis at Metropolia UAS, in collaboration with Superhood Oy.
Focus
Professional & Research Work
Internship – Document-Level Location Selection | Superhood Oy
Built a FastAPI microservice for Finnish news article processing, implementing NLP preprocessing and location extraction using Stanza and Geoapify across four pipeline configurations: baseline, postal-first ranking, and static and dynamic resolver modes.
Master's Thesis – A Hybrid Rule-based & LLM Approach
Evaluated the internship pipeline through controlled experiments comparing static, API-based, and MCP-assisted approaches for location disambiguation. Achieved 83.33% exact match accuracy across 60 annotated articles using a Groq-hosted LLM for context-aware ranking.
VoIP Backend – Asiakasgroup Oy | IMIB network
Designed REST endpoints and database models for a Flask and MySQL backend system supporting CRUD and bulk operations and integrations. Contributed to structured API design, database integration, and collaborative development within a production-oriented environment.
Public Demos
Machine Learning Experiments
Regression & Classification Experiments
Reproducible Python experiments including Ridge/Lasso regression, SVM, KNN, feature selection, and structured evaluation.
Skills
Languages: English (Fluent) · Finnish (Basic – A1/A2)
Contact
Email: irum.shehryar@gmail.com