Bioinformatics and AI MSc postgraduate personal statement example

Personal statement example

Last spring I left a cheap recorder on my garden fence for a week and ran the files through BirdNET, the free neural network tool that suggests species from audio. Most of its labels were sensible, but it repeatedly called a wren a goldcrest, and only on recordings made in the wind. Working out why became a small weekend project. I compared spectrograms, cut clips to remove low-frequency noise and found that the errors largely disappeared once the gusts were filtered out. It was a modest exercise, but it showed me something I want to study properly: a model's confidence depends on the data it was given, and understanding a biological signal is often the key to improving the computation applied to it.

My degree in Biological Sciences gave me a broad grounding in genetics, plant physiology and statistics. My strongest work came in my final year, when my project re-analysed a publicly available RNA-seq dataset of Arabidopsis thaliana under drought stress. I had no wet-lab component, so the challenge was learning to treat someone else's data carefully. I checked read quality, aligned reads to the reference genome and used DESeq2 in R to identify differentially expressed genes. My first results were dominated by a single sample, and a principal component analysis showed it sat apart from its replicates. Rather than simply removing it, I went back to the metadata and found it had been collected at a different time of day, which matters for a plant with strong circadian regulation of gene expression. I discussed this openly in my report and ran the analysis both with and without the sample. My supervisor commented that this section was the most convincing part of the project, and I enjoyed the reasoning more than any lecture-based module.

Since graduating I have worked as a science technician in a secondary school. Much of the job is practical: preparing solutions, maintaining microscopes, and setting out equipment for thirty students at a time. It has made me organised and calm under time pressure, since a Year 9 practical cannot wait for a missing set of agar plates. I also rebuilt the department's equipment inventory, moving it from three conflicting spreadsheets into one shared sheet with consistent naming and simple validation rules. Teachers now find what they need without asking me, which saved more time than I expected and taught me how much useful work lies in tidying data so that other people can trust it.

Outside work I have been teaching myself Python through online courses, working through the basics of pandas and scikit-learn by building a simple classifier on a public dataset of iris flower measurements, then on features extracted from my own bird recordings. I know these are introductory steps. They have shown me where my knowledge stops: I can train a model, but I want a firmer understanding of why particular architectures suit sequence or image data, how to validate them honestly, and how to avoid the overconfident conclusions I nearly drew from my outlying RNA-seq sample.

I also help my grandfather with his allotment on Sunday mornings, which has nothing to do with computing but has kept me interested in how plants respond to dry summers, the question my project began with.

I am applying for postgraduate study in bioinformatics and artificial intelligence because I want to combine the biological judgement my degree developed with rigorous computational training. I would particularly like to work on gene expression and plant stress data, where careful handling of noisy measurements could improve how we interpret crop responses to drought. I bring patience with messy data, practical reliability from my job, and a habit of questioning results before accepting them.