I am a Research Fellow at the University of Birmingham working in AI (artificial intelligence), data science and engineering.
I'm helping BTRU Researchers with the PRO-CAR-T study. They have developed a system to monitor CAR-T therapy - that includes a patient app. It would be really helpful if this system could warn patients and their healthcare teams about serious side-effects or complications of treatments - called adverse effects. To help - we are applying artificial intelligence to predict adverse events that cancer patients may experience within 28 days of receiving CAR-T therapy. We are analysing a large patient dataset that includes demographics (age, gender, ethnicity), clinical variables, treatment information and detailed adverse-event records for each patient. The eventual aim is to see if the patient app can provide early warnings of these events - so patients can get the treatment they need as soon as possible.
My other work involves predicting the best combination of blood-pressure medicines for people by learning from their individual characteristics—age, sex/gender, BMI, medical history, labs—and from metabolomics data (tiny molecules in blood—the body’s chemical traces of how people process food and medicines. It is kind of metabolic fingerprint that helps us anticipate which treatments are most likely to work safely for the patient).
What does a typical day involve?
My days are fast-paced because I’m involved in several projects. I start with reflection and prayer, then a gym session, and I’m at my desk by 9 a.m. with tea and a couple of biscuits. I set three clear priorities and structure the day around focused work and meetings. As I’m also teaching and tutoring this year, some days include student meetings and preparing course materials. I usually wrap up around 6–7 p.m. and unwind by cooking—often Mediterranean or Asian dishes. I stay organised and motivated by drawing inspiration from voices like Denzel Washington and Jim Rohn for having a healthy/organised/happy and successful life.
What do you like most about your job?
I love turning complex, messy health data into useful tools that help clinicians choose safer, more effective treatments. The best part is turning research into real world solutions: linking information from electronic health records, text, sensors, and metabolomics so that decisions take into account risks, treatment, and safety all at once. I value patient and public voices—learning from lived experience and integrating it (carefully and ethically) to prioritise interventions people will actually accept. Finally, teaching is a joy, bringing live research questions into class, seeing students build and consider models, and making the classroom inclusive and accessible for mixed backgrounds is delightful.
What made you decide to work in this area?
My background is in computer science, but I wanted work that connected algorithms to human language and impact. Being multilingual (Arabic and its dialects, French, English) - Natural Language Processing (NLP) was a perfect fit: it lets computers understand and learn from what people say. I began with sentiment analysis - a way for computers to figure out how people feel by detecting whether comments were positive or negative. From there I realised NLP and AI can do far more—anticipate health risks, suggest the best treatments, and highlight habits that might cause harm—so I focused my research on applications that improve care. That practical, people-centred impact is why I chose this area.
What is the hardest part of your job?
Knowing how much my work can shape people’s lives is the coolest part of the job—and exactly what makes it the toughest.
What is your greatest success at work to date?
My biggest successes span research and teaching. As a researcher, I’ve repeatedly started projects from scratch—like creating tools to understand emotions in Algerian language fro my PhD, building a real-world dataset from Scottish hospital records, and designing models to help choose the best blood pressure treatments based on lab results. As a teacher, the wins I’m proudest of are when difficult ideas “click” and students say the course has shaped their plans for the future.
Are there any aspects of your work that might surprise people?
People are often surprised that most of my time isn’t “building AI,” but getting permission to access data, cleaning messy discharge summaries and metabolomics files, and writing documents so everyone follows the rules to keep everything safe and legal. We don’t just predict who will benefit; we also try to foresee side-effects and fairness issues, then fine-tune and explain the model so recommendations are safe and equitable. And a surprising amount is human work: co-designing plans with clinicians and patients, translating results into plain English, and teaching so the tools are actually used.
Have you attended any interesting work-related events recently – or had some good news?
Yes—this year at the Association for Computational Linguistics (ACL) in Vienna I had my first paper accepted to the main conference. With an acceptance rate around 20–25%, it’s notoriously competitive, so this was a big milestone for me. I’ve published in several ACL workshops before (also selective), but making it into the main programme felt especially meaningful.
How do you think your work might change in the near future?
We’re moving from tools that do just one thing to smarter systems that don’t just predict health risks—but also suggest treatment options, explain why, track side effects, and update recommendations as things change. We’re combining different types of health data—like doctor’s notes, lab results, vital signs, metabolomics, and scans—to learn how treatments work together over time, not just one drug at a time. And we’re working to make these predictions more reliable, especially when there’s uncertainty.
How do you explain your job to other people and what do they normally say? I teach computers the way you’d teach a child: show lots of examples, let them learn the pattern, then test them on new, unseen cases. If they struggle, I refine the lessons and retrain; if they do well, we trust them with more cases. People usually say “Impressive”—sometimes because the idea is simple, sometimes because it sounds more mysterious than it is.
What do you enjoy outside of work?
Outside of work, I prioritise family—even though most are abroad. I make time for regular video calls and I travel mainly to see them. I also carve out quiet time for reflection (even if my thoughts drift to research, papers, and funding), and I’ve recently discovered I really enjoy cooking. To switch off, I’ll put on a Netflix movie and unwind.
What would your dream job be – if you were not a researcher/scientist?
If I weren’t a researcher/scientist, I’d still be in the classroom as a lecturer—that part of university life really energises me. I could also see myself running my own Research and Development-driven venture, turning ideas into useful products. The most unrelated path that genuinely tempts me is immigration law: after moving to the UK I became fascinated by the rules and routes, and I’d love to help people navigate them. Also, I’ve always wanted to visit Harvard Law School, inspired in part by my hobby of watching law series.
Which scientist or other person (living or dead) would you most like to meet?
This is a tricky one. I’m most inspired by people whose journey mirrors mine—much like an AI model matching a new patient to similar cases to find what works. I’d choose someone who shares my background—a woman, Muslim, from a minority group—who has faced similar challenges with the same resilience and values, and who is where I want to be: leading a business, sharing knowledge, inspiring others, building useful tools, helping people, and staying humble and family-oriented. Because everyone is unique, it’s hard to name a single person; in truth, every scientist and every person I meet shapes who I’ll be tomorrow, and as a naturally sociable person I’m happy to meet people in general rather than single out a name.