(Apologies for cross posting. This project also has open PhD positions.)


Postdoctoral fellow open in the "Democratising Big Machine Learning" project in the Department of Computing and Information Systems at the University of Melbourne, Australia.


The project is on machine learning and systems for data preparation. Topics span probabilistic databases, data integration/entity resolution, adaptive importance sampling for crowd sourcing, ML workflows – it is expected the candidate would work in one of these areas or related, but likely not all.


Ideal candidates would have strong background in some of: Bayesian inference, probabilistic graphical models, optimisation, linear algebra, mathematical statistics, experimental design, databases (the probabilistic kind and otherwise), cloud platforms like EC2/Azure. A strong mathematical background is a must. The project involves systems building; a major activity will be publishing.


The  project's sole-CI Ben Rubinstein is a junior faculty member in the department (Assistant Professor equivalent) with a PhD from Berkeley and 3yrs as Researcher at the Microsoft Research Silicon Valley lab. He has worked in many of the major industry research labs, and has served on the PCs of many of the major ML, AI, DB conferences. The group has strong connections and support from industry, specifically for this project (with significant time on Azure). The project is fully funded by the ARC (Australia's NSF). The broader group at Melbourne is very strong, eg hosting both SIGMOD and CIKM this year, and with ties with an exceptional mathematical statistics group at Melbourne lead by Peter Hall. The University of Melbourne is ranked as the top in Australia, and among the top 30-40 institutions internationally, while the city of Melbourne is consistently ranked among the top handful of most liveable cities for its exceptional quality of life.


The position is funded for 1 year with a competitive salary of $80k AUD plus 17% superannuation.


For more:

Email Ben Rubinstein [log in to unmask]