About

If nothing fits, I build my own.

That goes for the app I run my week on, and for the week itself. Something always needs to be happening: labs and workshops, the library after, runs in the park, coffee with someone, evenings perfecting what I’ve built. I do a lot, I do it my own way, and I make room for my health so it lasts.

A plane at the gate at night, its engine and wing lit by the apron lights

A rocky cove and beach in black and white

A grey sports car parked beneath a tall residential block

Looking down a rocky slope to a cove and the sea

Shot on iPhone 17 Pro.

Why drug safety

The scariest spike I ever found turned out to be a wave of lawsuits, not a new side effect. It moved a share price and a legal strategy long before it moved a safety conclusion — and separating those three is the whole job.

Most people in drug safety aren’t strong on data; most data people don’t know drug safety; and the commercial question is usually asked by a third group who talk to neither. I’m trying to be useful at the join.

I’m looking for a 2027/28 placement year in the pharma industry, in drug safety, real-world evidence or clinical data — the kind of work where the analysis has to hold up in front of a regulator, a court or a board.

Studying

MSci (Hons) Pharmacology, King’s College London

Based

London | Dubai

Year 2

Endocrinology, Drugs & Disease, Research Skills

The journey

How I got here.

2025

Couldn’t write a line of Python. Started learning data — in public.

Early 2026

First project: 20M+ adverse-event reports. Found a scary spike was litigation, not harm.

Spring 2026

SQL, patient records, and a signal detector validated on a known warfarin risk.

Mid 2026

ML, survival analysis in R, then AI reading medical text — the fanciest model came last.

Late 2026

Causal inference: a painkiller that looked twice as safe as it was, and two methods that caught it.

A 2027/28 placement year in pharma: the same questions, asked of real clinical and safety data.

What I think

Three things I’d argue for.

A rise in reported side effects is treated as a rise in harm. It routinely is not.

Adverse-event reporting is voluntary, so the count measures who is paying attention — media coverage, lawyers advertising, a regulator asking questions. When one heartburn drug’s reports exploded in 2020, the cause was litigation, not new clinical harm. The reports moved before the evidence did.

Evidence Twenty million FDA reports, and the Zantac timeline

What would change my mind A spike that holds up after adjusting for prescriptions dispensed and media volume would be a real signal, and I would want to see that adjustment before dismissing any of them.

Life sciences is buying AI on sophistication. Sophistication is not what predicts performance.

I put four systems on the same task of finding drug side effects in medical text. A model trained on biomedical literature won. A plain keyword matcher beat a general-purpose AI model. A general LLM with no task-specific training came last, and invented details that were not in the source. The ranking tracked fit to the domain, not capability or cost.

Evidence Four models, one test set, measured hallucination rate

What would change my mind This was a deliberately worst-case setup for the LLM — no fine-tuning, no examples. A few-shot version would likely close some of the gap, and I would want that comparison before generalising the claim to procurement.

When patients are not randomly assigned, the bias has a direction — and it flatters the drug.

Doctors steer frailer patients away from riskier drugs. That is good medicine and it quietly makes the risky drug look safe in the data afterwards. On a dataset where I had sealed the true answer in advance, the obvious comparison understated a real harm by 46%. Two independent methods recovered it. More data would not have helped: the problem was structural, not statistical.

Evidence A sealed true effect of 2.20, and two ways of finding it

What would change my mind The direction is not a law. Where the sicker patients get the newer drug, the same mechanism runs the other way and exaggerates harm instead — so the honest claim is that the bias is directional and knowable, not that it always favours the drug.

Education, experience and skills are on the CV