Are you inspired by invention? Is problem solving through teamwork in your DNA? Are you a scientist who is not only interested in the latest research in Machine Learning (ML) and Natural Language Processing (NLP), but also passionate about applying research in building solutions that enrich viewing experience of millions of customers? Answer yes to any of these and you've got to check us out!
Translation Service is a brand-new initiative to centralize and automate movie subtitle translation processes within Amazon Video. When launched, Translation Service will be an integrated, full-service, workflow for all forms of subtitle translation for all AV titles. It will have an integrated workflow that goes through extensive machine translation using core ML and NLP techniques for the final output.
This is an early stage initiative with a fast paced, highly collaborative start up like environment inside Amazon Video. To be successful you need to be flexible and entrepreneurial. You are a true owner - you are hands on and you lead by example. You acquire expertise as needed, iterate and adapt your approach as you learn more. You are deeply interested in pragmatic problem solving as well as taking risks in pioneering brand new areas. You pioneer, select and train appropriate machine learning models used in Machine Translations and NLP.
If you like to work hard, have fun, and make a difference, this is the team for you.
5+ years of relevant work experience.
A PhD or MS in CS / Computational Linguistics / Machine Learning / Deep Learning, Statistics, or in a highly quantitative field
Min. 2 years of experience in modelling / realizing algorithms in Machine Translation(MT) with Deep Learning.
Must have experience in designing & building working models - using statistical and neural techniques.
Work experience / Deep understating / Knowledge in NLP, MT – Rule based, Statistical, Hybrid, Neural.
Must have experience in building RNN (must), DNN, CNN, LSTM model and using commercial framework like TensorFlow/Caffe/Theano
Min. 1 year experience in coding (realization of models / PoCs / etc.) Java, C++, or other programming language, as well as with R, MATLAB, Python.
End to end work experience in MT ( problem definition, Data definition/ curation, metrics, choice of models, justification, model building, proof of concept, benchmarking, algorithm optimization, (iterations), & ship to product implementation )
Ability to define ground truth (data definition & rules), training & validation data sets
Excellent written and verbal communication skills, ability to communicate effectively to both technical and nontechnical audiences.
Excellent problem solving abilities
Publications at top-tier peer-reviewed conferences or journals
Proven track record of innovation in creating novel algorithms and advancing the state of the art (especially in MT )
Experience working with large datasets and Deep Learning algorithms