Ahmed Mustafa Younes
United Kingdom
Summary
Researcher and software engineer with a PhD in NLP, working at the intersection of research and building systems. Focused on evaluating and improving the interpretability of transformer-based large language models through an extensible diagnostic framework developed during my PhD. Experienced in researching and building end-to-end data pipelines for large-scale multilingual text, as well as interactive dashboards that translate research outputs into practical tools used by clients. I have also worked on evaluating the robustness of customer-service chatbots by developing semi-automated LLM-as-a-judge evaluation workflows to support faster iteration and production monitoring. Overall, my focus is on translating research into practical tools for clients and product teams.
Overview
6
6
years of professional experience
10
10
years of post-secondary education
Work history
Digital Learning Consultant
QA LTD
2025.08 - 2026.02
- Deliver postgraduate-level programmes in data science and AI, supporting learners in applying machine learning and NLP techniques to real-world datasets.
- Mentor learners on data preparation, exploratory analysis, and evaluation, with an emphasis on building reproducible and well-documented analytical workflows.
- Support applied projects involving text and numerical data, helping translate theoretical methods into working prototypes.
- Engage in ongoing professional development in cloud computing, machine learning foundations, and applied data science, incorporating new tools and practices into programme delivery.
- Collaborate with programme teams to align technical content with industry standards and responsible-AI practices.
Applied LLM Research
EMOTECH
2025.01 - 2025.03
- Worked on the evaluation of customer-service chatbots across Arabic dialects and English, focusing on response quality, task alignment, and robustness.
- Developed semi-automated evaluation workflows using an LLM-as-a-judge approach to reduce manual review and support faster iteration during development.
- Designed prompt-based test scenarios to probe failure cases such as hallucinations, brittle logic, and inconsistent behaviour across conversational contexts.
- Built lightweight evaluation pipelines to track model outputs, compare versions, and monitor behaviour over time.
- Collaborated with engineers and product stakeholders to use evaluation findings to inform system design and deployment decisions.
NLP Researcher
CASM TECHNOLOGY
2020.01 - 2024.12
- Designed and maintained end-to-end NLP pipelines for supervised and unsupervised analysis on large-scale textual data.
- Built data-processing workflows covering data collection, ingestion, transformation, and preparation of unstructured and semi-structured text.
- Developed semantic-mapping and representation-based analysis workflows using embedding models, clustering techniques, and Hugging Face models to support exploratory analysis, topic discovery, community detection, and segmentation of multilingual datasets, with a focus on social-media text.
- Applied contrastive learning and similarity-based retrieval techniques to guide representation spaces toward topics of interest, enabling unsupervised exploration and weakly supervised classification, and improving clustering coherence and semantic search on multilingual data.
- Fine-tuned and built evaluation pipelines for transformer models across tasks including named entity recognition, machine translation, text classification, zero-shot learning, and natural language inference, covering up to 50 languages.
- Delivered interactive dashboards and analytical reports for external clients, translating technical results into interpretable outputs and iterating based on stakeholder feedback.
- Worked closely with data engineers, software engineers, researchers, and domain experts to troubleshoot pipelines, document workflows, and support reproducible data-curation practices across teams.
Education
PhD - Natural Language Processing
UNIVERSITY OF SUSSEX
United Kingdom 2020.01 - 2025.01
MSc - Data Science
UNIVERSITY OF SUSSEX
Brighton 2018.01 - 2019.01
Bachelor - Computer Science
HELWAN UNIVERSITY
Egypt 2012.01 - 2016.01
Skills
Programming & Tools
Python, SQL, Hugging Face, Transformer-based models, API-based LLMs, Docker
Machine Learning & NLP
Natural Language Processing, Multilingual NLP, Representation Learning, Transformer Models, Named Entity Recognition, Text Classification, Machine Translation, Natural Language Inference, Zero-shot Learning
Model Evaluation & Analysis
Model Evaluation, Diagnostic Analysis, Representation Analysis, Robustness Testing, Error Analysis, Interpretability, LLM-as-a-Judge Evaluation
Data & Pipelines
Large-scale Text Processing, Semantic Mapping, Clustering, Embedding-based Retrieval, Data Annotation Workflows, Reproducible Pipelines
Visualisation & Reporting
Interactive Dashboards, Analytical Reporting, Visual Analytics
Timeline
Digital Learning Consultant
QA LTD
2025.08 - 2026.02
Applied LLM Research
EMOTECH
2025.01 - 2025.03
NLP Researcher
CASM TECHNOLOGY
2020.01 - 2024.12
PhD - Natural Language Processing
UNIVERSITY OF SUSSEX
2020.01 - 2025.01
MSc - Data Science
UNIVERSITY OF SUSSEX
2018.01 - 2019.01
Bachelor - Computer Science
HELWAN UNIVERSITY
2012.01 - 2016.01