Postdocs in Machine Translation at the University of Edinburgh

Deadlines: 14 May (TRAIN@Ed) and 29 May, both 5 pm UK time.

At the University of Edinburgh, I am hiring Research Associates (postdocs) to work on machine translation. The following areas are particularly relevant, but candidates will be considered in all areas of machine translation:

Depending on your interests and eligibility, you might be funded by: The starting salary is £33,199-£39,609. The university provides visa sponsorship including reimbursement of standard fees.

The Positions

There are currently two job applications open. Apply to both of them if you can. Sorry for the extra bureaucracy.

One position is through the TRAIN@Ed program. This lasts a full 3 years, includes £5000 in relocation, and has a few other benefits. It is funded partly by Marie Skłodowska-Curie actions so there are eligibility requirements and a more complicated application process. Don't worry about having a company on board; contact me (jobs at and I will match your interests with some of the companies I work with. Applications for this position close on 14 May at 5 pm UK time.

The second position is funded by my grants. There are fewer eligibility requirements and there's no 5-page research statement to write. However, you won't get relocation and the current funding ends December 2021. The start date is negotiable. Applications for this position close on 29 May at 5 pm UK time.

Currently both of the above positions are targeted at people with a PhD, soon to be awarded a PhD, or equivalent experience. If you're interested in working on ParaCrawl or low-resource MT, contact me about a BSc/MSc-level position coming soon. Or you can apply to do a PhD with me.


Both positions have specific qualifications in the job ad. I'm looking for NLP (and ideally translation) research experience or strong core computer science skills (data structures, C++). Nobody matches exactly.

About the group

The University of Edinburgh hosts one of the largest translation research groups in the world. We are part of 5 of the 7 Horizon 2020 Multilingual Next Generation Internet projects. The group owns a cluster with 80 GPUs and 1.2 PB space. We also have 300,000 GPU hours on Wilkes2 to burn in one year.