Omar Eissa

Data analyst and Computer Science graduate (First Class Honours, 86%). I check whether the number is right before anyone acts on it: I have rebuilt a production pipeline that was quietly producing wrong outputs, caught a methodological flaw in my own analysis partway through, and reported a forecast of mine that ran 24% low rather than presenting a cleaner one. Yearlong placement as the data engineer on a machine learning research study, plus a commercial internship in credit risk at a retail bank.

I write the SQL, clean what comes back, and present the answer to people who do not want the technical detail.

Research Assistant, Data and Analysis at AI4SW (Artificial Intelligence for Student Wellbeing), University of Huddersfield (2024-07 – 2025-07)

Owned the data engineering for a research study predicting student wellbeing, where participants contributed wearable health data and daily mood check-ins through a React Native mobile app.

• Inherited the Firebase to CSV pipeline extracting that data producing systematically inaccurate outputs that had gone unnoticed, because the numbers looked plausible. Traced the fault to root cause, proved the mechanism, and rebuilt the pipeline. The corrected data now supports ongoing published research and I am credited in the forthcoming paper; in its previous state it would not have got there.
• Converted multi year Apple Health XML exports and Firebase JSON logs into structured daily and hourly tables, reconciling incompatible export formats and irregular participant compliance across a three month collection window so the study had a dataset that could actually be modelled.
• Built the feature layer joining heart rate, sleep and activity signals against daily mood check-ins, including the anomaly detection that flagged device non-wear periods which would otherwise have been modelled as genuine low activity days.
• Built the weekly Excel and Power BI reporting the study ran on, taking progress, findings and data quality issues to the research supervisor and shaping the direction and methodology the project took.

Credit Risk Modelling Intern at Credit and Risk Management Division, Credit Agricole Egypt (2023-08 – 2023-08)

• Segmented three years of auto loan portfolio data into high, medium and low risk categories using Python (pandas, NumPy, Matplotlib), and built the Excel summary tables and visuals used to present the findings directly to senior management. Completed structured Python data analytics training alongside the risk modelling team.
• Defended a segment carrying materially higher non-payment risk than the team expected, which meant showing the pattern held rather than being an artefact of where the segment boundaries had been drawn.

Retail and Events Team Member at University of Huddersfield Students' Union (2025-08 – Present)

Customer facing role in a high volume live events and retail environment, held alongside full time study.

BSc (Hons) in Computer Science – University of Huddersfield (2022-09 – 2026-07)