Open to opportunities

Khalil Alakbarzade

@khalilalakbarzade

Message

I build end-to-end AI and machine-learning solutions that turn complex data into business impact.

United Kingdom

Message

What I'm looking for

I'm looking to build end-to-end AI and machine-learning solutions that turn complex data into measurable business impact, particularly across NLP, fraud detection, anomaly detection and data-driven products.

I'm leading LibrAI at the University of Sheffield Library, where my team won First Prize and the People’s Choice Award at the 2026 Data & Design Hackathon. We built an AI video-intelligence platform that makes university video content searchable and accessible, saving over £14,500 of manual work.

I've built Python NLP pipelines using OpenAI Whisper, automated subtitles, summarisation, entity extraction and topic classification, alongside Flask services for video ingestion and search. I’m also building SIM-box telecom fraud detection with Smartel using live CRM data, time-series anomaly detection and transformers, supported by projects in recommendation systems, forecasting, classification, financial analysis and large-scale PySpark analysis.

Experience

Work history, roles, and key accomplishments

Current

Data Science Team Lead

Current

LibrAI, The University of Sheffield Library

Mar 2026 - Present (5 months)

Led a 3-person team to First Prize and People's Choice at the University of Sheffield Data & Design Hackathon 2026 by building an AI video-intelligence platform. Selected by the University Library to take the winning prototype into production, leading 3 engineers on scope, delivery, and code quality.

Python NLP Flask JSON Full Text Search Team Leadership

Current

Data Scientist

Current

University of Sheffield (with Smartel)

Jan 2026 - Present (7 months)

Building a SIM-box fraud-detection system on Smartel's live CRM and telecom data, combining time-series anomaly detection with transformers to flag fraudulent SIM activity. Engineering temporal features from call-detail and subscriber records and benchmarking anomaly-detection models to maximize fraud recall while controlling false positives.

Python Anomaly Detection Transformers Feature Engineering Fraud Detection Benchmarking