New Opportunities

I was unsucessful in applying for a permanent position at QUT at the end of last year, so started looking at positions at other universities in Brisbane, at other universities in Australia, and at other universities around the world.

After a few rounds of waiting to hear back, and waiting, and being unsuccessful, and waiting, I had two interview requests occur at the same time for some positions in the UK. After a bit of back and forth about how to attend an interview on the other side of the world (hint: it’s Skype, unless you’re meant to do two days of interviews, meet and greet and seminars), I have been offered and accepted a position with the London School of Hygiene and Tropical Medicine for statistical and machine learning work on epidemiology data, looking at vaccine strategies.

It’s been really interesting working with my QIMR colleagues on the dengue project, and I’ve certainly been pushed to pick up more skills when it comes to parameter estimation in ODE models (which is sometimes known as model calibration), have learned a little molecular biology, and have got more experience working with a different type of scientist. Extending this to work on vaccine models and further epidemiology (Flasche et al. 2017) is going to be really exciting, and I’ll be picking up more machine learning skills in the process.

Part of the role involves teaching into some of the Masters courses; LSHTM in fact has no undergrad students As much as I’ve enjoyed teaching first year maths and stats to science students, teaching to those choosing to enrol in a Masters programme or undergoing continuing professional development is something I’d like to get more experience with. A few years ago I taught some colleagues of mine within CAR about Bayesian modelling of epidemiology data with JAGS, and talking through some statistics ideas with the QIMR folk has been worthwhile. I’ve also got the chance to develop some teaching material as part of a webinar series for mathematics for machine learning, looking at how concepts from calculus, linear algebra and statistics are used in machine learning.

This may be the push I need to learn a bit more about python, which seems to be the standard approach to problem solving in ML. Packages like pytorch, pandas, or scikit-learn seem to have a lot going for them, and it strikes me that while I can’t follow every trend, not adapting with the field is a good way to end up without the skills that are valued. I’ve been meaning to look into Statistical Rethinking (McElreath 2018) and there’s a jupyter notebook version of the code for once I’ve finished learning more about how to use rstan.

There’s a little bit more work to be done on the UPTECH project, adapting some of the methodology of Clifford et al. (2018) to look at inflammation of the peripheral airways.

So yes, lots to learn and lots to organise in terms of moving to another country.

References

Clifford, Sam, Mandana Mazaheri, Farhad Salimi, Wafaa Nabil Ezz, Bijan Yeganeh, Samantha Low Choy, Katy Walker, Kerrie Mengersen, Guy B. Marks, and Lidia Morawska. 2018. “Effects of Exposure to Ambient Ultrafine Particles on Respiratory Health and Systemic Inflammation in Children.” Environment International 114:167–80.

Flasche, Stefan, John Ojal, Olivier Le Polain de Waroux, Mark Otiende, Katherine L. O’Brien, Moses Kiti, D. James Nokes, W. John Edmunds, and J. Anthony G. Scott. 2017. “Assessing the Efficiency of Catch-up Campaigns for the Introduction of Pneumococcal Conjugate Vaccine: A Modelling Study Based on Data from Pcv10 Introduction in Kilifi, Kenya.” BMC Medicine 15 (1):113. https://doi.org/10.1186/s12916-017-0882-9.

McElreath, R. 2018. Statistical Rethinking: A Bayesian Course with Examples in R and Stan. Chapman & Hall/Crc Texts in Statistical Science. CRC Press. https://books.google.com.au/books?id=T3FQDwAAQBAJ.

comments powered by Disqus