Connectionists: JOB: Machine Learning and Speech Recognition Developers

[Tony Robinson]

Who'd have thought that two decades later Recurrent Neural Nets would 
return to become the dominant means of doing Speech Recognition.

More details at https://www.speechmatics.com/careers or by email - some 
results on huge RNNs for language modelling are to be presented at 
ICASSP15 and can be found at http://arxiv.org/abs/1502.00512.


Tony

MACHINE LEARNING & SPEECH RECOGNITION DEVELOPERS

Location: Cambridge, UK
Contact: careers at speechmatics.com

Background:

Speechmatics provides some of the world’s fastest and most accurate 
speech recognition technologies, based in the cloud (SaaS).  We are a 
Cambridge-based company that started commercialisation last summer and 
already support several large customers as well as many small 
companies.  We are expanding rapidly and are seeking another 10-15 
people in the coming months. This is an opportunity to join a high 
growth team and form a major part of its future direction.

The Opportunity:

We are creating ground-breaking technologies in the fields of machine 
learning, big data and speech recognition and as a result are looking 
for several machine learning & speech recognition developers. Your role 
will be to develop disruptive technologies in one of the following main 
speech recognition areas: decoding, acoustic modelling and language 
modelling. You will spend the majority of your time developing and 
improving one of these areas as well as collaborating closely with other 
team members to build large scale systems.  Much of our work is 
cutting-edge research thus we are seeking individuals who are motivated 
by developing disruptive and novel approaches to machine learning and 
speech recognition. This role offers challenges across a wide variety of 
projects and responsibilities, including the opportunity to influence 
the direction of the business.

Qualification:

To succeed in this role, you must be a strong programmer and creative 
problem-solver who thrives in fast-paced environment. Your background is 
a numerical degree (2.1 minimum) and ideally practical experience in 
either big data, machine learning and/ or automatic speech
recognition.  You should also have some of the following skills:

Essential:

* Strong programming skills in at least Python or C/ C++, additional 
languages a bonus
* Excellent academic record of achievement, having achieved a 2.1 or 
above in Computer Science or other similar technical/ scientific discipline
* Strong Linux scripting skills
* Theoretical Machine learning knowledge (with at least basic knowledge 
of modern ASR)

Desired:

* Experience with building machine learning libraries
* Experience building and training neural networks (RNN/ DNN/ LSTM)
* Experience with Kaldi, HTK or similar
* Experience writing efficient CUDA code
* Ability to implement low level code optimisations

Preferable:

* Implemented ASR on Android and/or iOS or experience programming 
smartphones
* Knowledge of WFST / lattice processing algorithms
* Worked with reverberant and noise robust speech recognition
* Implemented diarisation and/or speaker identification

We are building platform technology and therefore are challenged 
continually by the new uses that our customers are finding for it. The 
ability to take up a new challenge, quickly turn it around using an 
automated/machine learning solution and to personally learn from the 
experience is core to our team culture.

Salary:

We offer a competitive salary, bonus scheme and a generous EMI share 
option scheme.

-- 
Dr A J Robinson, Founder, Cantab Research Ltd
Speechmatics is a trading name of Cantab Research Limited
We are hiring: www.speechmatics.com/careers 
<https=www.speechmatics.com/careers>
Phone direct: 01223 778240 office: 01223 794497
Company reg no GB 05697423, VAT reg no 925606030
51 Canterbury Street, Cambridge, CB4 3QG, UK
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/connectionists/attachments/20150411/2b113bb1/attachment.html>