You're here for a reason. I'm Sepehr Zarnegar.

Explore research View CV
Sepehr Zarnegar in professional attire
Based inLondon, UK

01 / Research profile

Research
profile

I am a Warwick Business School graduate interested in how artificial intelligence changes organisational decisions, governance and work.

My experience combines qualitative research on AI governance and organisational adoption with quantitative projects that examine forecasting, data leakage and the limits of prediction.

Emerging research interests

AI governance & accountabilityHuman–AI decision-makingAI adoption & organisational change

02 / Academic journey

My journey

Education, research and applied experience, brought together as one continuous story.

Future

Next Chapter

What's Next?

September 2024 – September 2025

Education

Warwick Business School

MSc Management of Information Systems and Digital Innovation

University of Warwick Departmental Scholarship · Distinction in Research Methods. Modules included Data Analytics & Artificial Intelligence and Programming for Data Analysis.
2025

Dissertation research

AI accountability

From transparency to enforceability

Reflexive thematic analysis of regulatory texts, governance standards and corporate policies; development of an interoperability map and diagnostic tools for AI assurance.
May – July 2025

Qualitative research

Allnex

International TeamWork Programme

Warwick and Monash Universities · Nine semi-structured stakeholder interviews exploring AI adoption and knowledge management.
September 2019 – August 2023

Education

University of Tehran

BSc Industrial Management

17.43/20 · Top 10% of cohort · Four-year scholarship.
September 2019 – August 2023

Teaching

University of Tehran

Teaching Assistant · Strategic Management

Supported seminar preparation, case discussions and individual student learning; helped refine teaching materials and resources.
May – October 2023

Applied experience

Business analysis

BAT Pars & Eastcool

Consumer research involving 200+ respondents, competitive intelligence across 10+ brands, and experience with reporting automation and demand forecasting.

I didn’t come this far to only come this far.

03 / Selected research

Selected
research

GitHub ↗

MSc dissertation research / 01

Accountability for Agentic AI

How can organisations make accountability enforceable when workplace decisions are delegated to AI? My dissertation examined the relationship between transparency, institutional authority and meaningful human control.

2015–2025document scope2accountability dimensions
Reflexive thematic analysisDocument analysisAI governance
Read research summary ↗
Conceptual synthesis from the dissertation: answerability and enforceability underpin AI accountability, supported by evidence and meaningful human control.

Dissertation framework · conceptual synthesis

Organisational adoption study / 02

Understanding AI Adoption

What prevents employees from engaging with a corporate data and AI knowledge platform? In a Warwick–Monash project for Allnex, I conducted qualitative research into adoption barriers and internal communication.

9stakeholder interviews4adoption friction points
Semi-structured interviewsThematic analysisKnowledge management
Read study summary ↗
Four adoption barriers identified in the Allnex study: technical and UX accessibility, community isolation, content fatigue, and limited multi-channel visibility.

Qualitative findings · Allnex project

Quantitative research project / 03

After the Bell

Do Bitcoin’s weekend price movements help explain the next opening gap in crypto-linked equities? I studied 156 weekends using chronological forecasting, benchmark comparisons and explicit checks for information leakage.

156weekends studied3crypto-linked equities
Python & SQLExpanding-window OLSBlock bootstrapOut-of-sample evaluation
Read methods & scope ↗Live research demo ↗GitHub ↗
Mean absolute opening-gap error in percentage points for Bitcoin-linked versus historical-mean forecasts: COIN 1.314 versus 2.540; MSTR 1.288 versus 3.085; MARA 1.377 versus 3.046. Lower is better.

Historical research · predictive association

Machine-learning research audit / 04

E-commerce Purchase Audit

When should a predictive model not be deployed? I audited an e-commerce dataset, identified post-outcome leakage and measurement limits, and reported a negative test result rather than recommending an unreliable model.

0.468test ROC–AUC3model families evaluated
Logistic regressionRandom forestGradient boostingMeasurement validity
Read audit summary ↗GitHub ↗
Chart comparing the selected purchase model with a no-skill baseline, showing no improvement on the test set.

Negative result · deployment not supported

04 / Preparation for doctoral study

My research
toolkit

Qualitative inquiry

Semi-structured interviews, thematic analysis, reflexive document analysis and the interpretation of regulatory and organisational texts.

Quantitative analysis

Regression, predictive modelling, cross-validation, chronological evaluation, benchmark design and bootstrap uncertainty estimation.

Technical practice

Python, SQL, Excel, Tableau and Power BI. Reproducible workflows, data validation and attention to measurement and leakage.

Academic preparation

Distinction in Research Methods at Warwick Business School. GRE Quantitative: 170 · Verbal: 161 · Analytical Writing: 3.5.

Recognition & academic community

Warwick Departmental Scholarship (2024), Warwick Award and Warwick Volunteer Award (2025), and Google Data Analytics Professional Certificate (2026). Teaching assistance in Strategic Management and leadership as Head of Student Union at the University of Tehran.

05 / Contact

Let'sconnect.