Friday, March 10, 2017

Rise of the computationalists

As part of the Medicinal Chemistry List*, I have noticed how many computational positions in drug discovery I am running into. The latest one that caught my eye is this one from Pfizer Cambridge:
The Pfizer Computational Sciences group has an opening for a data scientist with expertise in machine learning. The successful candidate will work in close collaboration with groups across Medicinal Sciences organization and will utilize his/her machine learning, data analysis, and scientific programming experience to address challenging problems covering a wide range of research and development activities within Pfizer R&D. To be successful in this role, the incumbent must have the talent and skills to analyze large, multi-dimensional datasets from internal and external sources and to rapidly develop effective in silico models and implement powerful computational solutions.
  • Ph.D. in computational chemistry, computer science, physical or biological sciences, machine learning, or related discipline with 0-3 years of relevant experience required.
  • Familiarity with several machine learning algorithms and packages (e.g. Regression and Classification algorithms, Supervised and Unsupervised learning algorithms, Random Forest, Support Vector Machine, Neural Networks, Deep Learning, Sci-kit Learn, R, MATLAB, Theano, TensorFlow).
  • Experience working with large data sets, preferably in drug discovery setting.
  • Experience with Unix/Linux, HPC environments, and high-level programming language (e.g. Python).
  • Demonstrated track record of applied machine learning and data science through publications in top tier peer-reviewed journals and/or presentations in national or international conferences.
I find it interesting that this position is advertised at the "Principal Scientist" level (my memory might be fuzzy, but that's a R6 (thanks, Anon!) R5, right?) for a Ph.D. with zero years experience. (Typical entry-level Ph.D. is R4.) 

A long time ago, the calculation was that a single synthetic chemist ran $250,000 in terms of salary and overhead (this number must have changed by now) - I wonder how much a computationalist runs, in terms of overhead? More or less? 

*Yes, working on the process chemistry one. 


  1. Principal Scientist =R5 (at Pfizer), one step up from typicalentry-level PhD.

    1. Wasn't there an associate principal scientist level? I forget.

    2. No. There's an associate research fellow (R7) though.

    3. My guess is that the R4 starting salary isn't competitive in Cambridge given the experience they're looking for.