Advanced Technology Applications Center (ATAC), Global Databases Division, Global Infrastructure Sector
Grade - P3
Contract Duration - 12 months
Duty Station: CH-Geneva
Publication Date: 28-Feb-2019
Application Deadline: 22-Mar-2019, 12:59:00 AM
IMPORTANT NOTICE REGARDING APPLICATION DEADLINE: Please note that the deadline for applications is indicated in local time as per the time zone of the applicant’s location.
1. Organizational Context
a. Organizational Setting
The position is located in the Advanced Technology Applications Center (ATAC) of the Global Database Division, Global Infrastructure Sector. The Section is responsible for providing support in machine learning applications to other database sections, other WIPO divisions and other external entities. The Section has developed its own machine learning automatic translation tool, WIPO Translate for WIPO and to other international organizations. ATAC wishes to promote WIPO Translate by offering a software as a service solution to private and semi-private entities. A similar business model might be developed in the near future for WIPO speech recognition, image classification, chatbots and other ATAC tools.
b. Purpose Statement
The incumbent is responsible for insuring that the tools are of competitive quality and offering technical support to WIPO units and external Organizations. The support includes installing, maintaining and retraining running applications, usually based on machine learning software. The incumbent is also responsible for building a technical architecture which will be robust enough to be universally deployed with minimum maintenance cost effort. In parallel, the incumbent will explore new research algorithms (mostly using deep supervised machine learning techniques) that can be applied to improve WIPO processes and/or generate new business opportunities. The current exploratory work is divided into three main components: Text, Speech and Images. Text includes Machine Translation, text classification and chatbots; Speech includes speech recognition and speech translation; Image includes image classification and similarity search. WIPO Translate: WIPO?s Machine Translation engine relies on data-driven techniques, especially Neural Machine Translation (NMT). WIPO Translate, an in-house developed tool, is based on open source software (e.g. Marian).. A speech-to-text system is under development. The incumbent will work on organizing the workflow for training the speech to text and/or deploying in difference places. A product, based on machine learning, has been developed to automatically classify images. In the future, WIPO can choose to share this tool with external entities, in which case, the incumbent might have to work on packaging this tool in a way that render its use easier for intellectual property offices.
c. Reporting Lines
The incumbent reports to the Senior Machine Learning Specialist in the Advanced Technologies Applications Center.
2. Duties and Responsibilities
The incumbent will perform the following principal duties:
a. Take responsibility for workflow design and maintenance of WIPO Translate, master the current architecture and tools, explore hardware and software components that will provide the best price-efficiency model for training and decoding in various contexts, study solutions such as docker, cloud, GPUs, and others with a special focus on the adaptability/integration of the tool.
b. Stay abreast of Neural Machine Translation techniques, study in depth the supervised deep machine learning techniques, follow the state of the art to be able to experiment with next generation MT, define a strategy for experimenting with one or a few technology(ies) in our context, experiment and evaluate with various techniques for setting the best parameters for specific language pairs, specific domains, and for combining various input source corpora by mixing in-domain small corpora with larger out-of-domain corpora and by mixing multi languages in a single NMT model.
c. Manage MT-integration/ Natural language processing to build/maintain tools able to translate full documents e.g. texts, documents and HTML pages, design APIs/GUIs to improve means for accessing the output of machine translation
d. Develop methods to collect clean parallel patent sentences, e.g.: filter/clean titles and abstracts, align full texts using patent priority data, etc.
e. Define the training/development/test data sets, develop cleaning and filter methods on the data and perform quantitative and qualitative analysis
f. Work on automatic post-editing tasks: learn recurrent errors from human post editions, or from user feedback to correct MT output
g. Define a workflow for updating NMT models using newly published documents e.g. incremental training and explore online learning algorithms
h. Collaborate with various WIPO units and external entities to define possible areas of application of new and existing machine learning tools e.g. for text/image classification, speech recognition etc. and advise the team on ways to improve WIPO competitive advantage in AI.
i. Perform other related tasks as required
First-level university degree in Information Technology. An advanced university degree in Information Technology may substitute two years’ of required experience.
At least six years’ experience in machine learning related software engineering and/or computational linguistics.
Experience in using and designing neural networks (CNNs, RNNs, transformer models etc.), knowledge of Neural Networks frameworks: (Tensorflow, Pytorch, Theano, Keras).
Experience with machine translation (especially neural machine translation).
Experience in implementing Machine Learning projects and/or making significant scientific contribution in Machine Learning.
Excellent knowledge of written and spoken English or excellent knowledge of written and spoken French and good knowledge of English.
Excellent knowledge of English
Knowledge of other official UN languages.
Job Related Competencies (Essential)
Knowledge of machine learning techniques (especially neural networks)
Strong knowledge of programming language(s) is required (Java and/or Python and/or Perl and/or C++).
Knowledge of Unix
Good analytical and problem-solving skills and the ability to understand and document IT processes.
Excellent interpersonal skills and ability to maintain effective partnerships and working relations in a multi-cultural environment with sensitivity and respect for diversity.
Job Related Competencies (Desirable)
Statistics: automatic document classification approaches (SVN, Knn, EM, ANNs, Naive Bayes)
Databases: nosql techniques, Mysql, Oracle
Scripting languages: Perl, Python, bash
Unix: Ubuntu, Red Hat, configuring Unix remote servers (using command line mode)
Knowledge of version control systems (SVN and/or GIT) and deployment strategies (virtual machines, docker, cloud servers etc.)
Ability to write user guide, administration documentation and reports in English.
Excellent decision-making and problem-solving skills.
Knowledge of IP information tools, products and standards.
4. Organizational Competencies
1. Communicating effectively.
2. Showing team spirit.
3. Demonstrating integrity.
4. Valuing diversity.
5. Producing results.
6. Showing service orientation.
7. Seeing the big picture.
8. Seeking change and innovation.
9. Developing yourself and others.
Total annual salary consists of a net annual salary (net of taxes and before medical insurance and pension fund deductions) in US dollars and a post adjustment. The post adjustment (cost of living allowance) is variable and subject to change without notice in accordance with the rates as set within the UN Common System for salaries and allowances. The figures quoted below are based on the February 2019 rate of 67.7%.
Salaries and allowances are paid in Swiss francs at the official rate of exchange of the United Nations (not applicable for the External Offices).
Please refer to WIPO’s Staff Regulation and Rules for detailed information concerning salaries, benefits and allowances.
Temporary appointments are renewable, subject to continuing needs, availability of budget and satisfactory performance with a maximum cumulative length of two years.
Applications from qualified women as well as from qualified nationals of unrepresented Member States of WIPO and underrepresented geographical regions are encouraged. Please click on the following links for the list of unrepresented Member States and the list of underrepresented regions and the WIPO Member States in these regions.
The Organization reserves the right to make an appointment at a grade lower than that advertised.
By completing an application, candidates understand that any willful misrepresentation made on this web site, or on any other documents submitted to WIPO during the application, may result in disqualification from the recruitment process, or termination of employment with WIPO at a later date, if that employment resulted from such willful misrepresentations.
In the event that your candidature is shortlisted, you will be required to provide, in advance, a scanned copy of an identification and of the degree(s)/diploma(s)/certificate(s) required for this position. WIPO only considers higher educational qualifications obtained from an institution accredited/recognized in the World Higher Education Database (WHED), a list updated by the International Association of Universities (IAU) / United Nations Educational, Scientific and Cultural Organization (UNESCO). The list can be accessed through the link: http://www.whed.net/. Some professional certificates may not appear in the WHED and these will be reviewed individually.
Additional testing/interviewing may be used as a form of screening. Initial appointment is subject to satisfactory professional references.
Additional background checks may be required.