Bioinformatics · Brisbane, Australia
Antoine Haddad
Computational statistics · Multivariate analysis · Machine learning in biology
Biomedical science graduate completing a Graduate Diploma in Bioinformatics at The University of Queensland, with High Distinctions in computational statistics and in machine learning applied to biological data. I'm looking for work where biology, code and statistics meet.
About
I work in the space where biology, code and data meet. My background is biomedical — microbiology and immunology, molecular cell biology, physiology and pharmacology — and I moved into bioinformatics to work on the analysis side of that biology rather than only the bench side.
What I care about is staying honest about what the numbers actually support. Most of the interesting work in an analysis is deciding how much of the structure you found is real, and being willing to report a negative result when that's what the data says.
Alongside study I've held clinical administration work in a NATA-accredited respiratory and sleep practice, where accuracy in patient records and documentation was the whole job — useful preparation for research data that has to be right rather than merely finished.
Education
Graduate Diploma in Bioinformatics
Completed: Applications of Computational Statistics (High Distinction); Multivariate Data Analysis and Machine Learning in Biology (High Distinction); Bioinformatics 1 (Distinction); Mathematics for Data Science 1.
In progress: Bioinformatics 2 — Development and Research; Advanced Genome Informatics; Statistical Analysis of Genetic Data; Introduction to Programming.
Bachelor of Science, Biomedical Science
Microbiology & Immunology; Integrative Physiology & Pathophysiology; Molecular Cell Biology; Principles of Pharmacology; Analysis of Scientific Data.
Technical skills
Tools I've actually used on real data and can talk through in detail:
Projects
Testing for cryptic species in a coral gene-expression survey
A set of reef coral colonies had been assigned to a single species on morphology alone. Given a baseline gene-expression matrix — a few hundred colonies across several hundred loci — the question was whether that assumption survived the molecular data, or whether the profiles broke into discrete groups consistent with cryptic species hiding inside one morphological description.
I standardised the expression variables so no gene dominated on scale alone, screened for multivariate outliers using Mahalanobis distance on the leading components, then ran PCA and classical MDS as two independent views of the same variation. Rather than eyeball the ordinations for groups, I tested for them: k-means across a range of k, judged on within-cluster sum of squares and average silhouette width, and cross-checked against Ward hierarchical clustering. Sampling depth was held back and used only afterwards, as an ecological sanity check on whatever the clustering found.
The evidence pointed the other way. No component dominated the variance, both ordinations showed one continuous cloud rather than separated ones, and cluster support stayed near zero at every k — with the hierarchical solution agreeing. The defensible conclusion was a negative one: these data don't support a cryptic-species reading, and the report said so and set out what evidence would actually settle it. The real work was declining a clustering result that k-means will hand you whether or not the structure is there.
Reversi engine in Python
A complete implementation of Reversi built from a specification — board state, move legality, line-flipping in eight directions, turn handling and end-game detection. I wrote it as small testable units with their own test files rather than one monolithic loop, which is the habit that carries over into analysis code.
I also design and ship Obsidian plugins in TypeScript — four of them, released and in daily use. They're described on the other page since they're built for my own workflow rather than for research, but they're the largest body of code I've written and maintained.
Experience
Medical Administration Officer
Maintained referral and patient identity accuracy across multiple specialist divisions, collated clinical documentation from GPs, pathology and sleep services to deadline, and coordinated sleep-study and lung-function bookings. Contributed to clinic audits under the practice's quality-improvement program and presented status at quarterly meetings.
Traffic Controller
Directed pedestrian and vehicle movement on active worksites under strict safety procedure, across rostered weeks averaging 50 hours .
Kitesurf Instructor
Taught beginners to IKO methodology, including students with limited English, using demonstration-led instruction to hold safety standards in a dynamic water environment.
Student Leader & College Photographer
Elected to the student leadership team and served as college photographer, producing photography and content for college marketing to brief and deadline.
Retail & Customer Assistant
Customer service alongside inventory management and preparing stock for shipping.
Contact
The best way to reach me is email. Happy to talk about bioinformatics work, research assistant positions, or anything statistical.