Machine Learning Engineer: Speech Processing


Cambridge or Bristol, GB

Category:  Assessment and Research

Job Title: Machine Learning Engineer: Speech Processing

Contract: Permanent

Salary: £60k - £84k

Location: Cambridge or Bristol


Help us revolutionise English language learning and assessment!


We are recruiting a Machine Learning Engineer: Speech Processing to expand our in-house expertise in speech processing, to increase the pace and scope of our tech transfer efforts at English Language iTutoring (ELiT), which turn cutting-edge research into products and services that help people learn English.


A new team of Machine Learning Engineers will work closely with education researchers and with the  Speech Research Group at the Cambridge University Department of Engineering to develop existing technologies and create new ones.

Our ML-powered capabilities use state-of-the-art techniques to automate assessment and feedback on English proficiency. These technologies were developed with the Institute for Automated Language Teaching and Assessment (ALTA) and Cambridge Assessment English’s Research and Thought Leadership department. We work with multidisciplinary teams at the University of Cambridge and are looking for people who will enhance these collaborations.


This is an exciting opportunity to work with world-leading speech processing researchers, solving educational problems with machine learning and artificial intelligence.


What you’ll be doing

Developing a technical roadmap for an area where machine learning-based systems are applied. You will grow the machine learning and natural language processing capabilities of our organisations.


Leading on collaborative technology transfer projects with product teams, academic researchers and consultant engineers. This will include prototyping, operationalising and deploying novel machine learning developments to solve business problems.


Collaborating with data engineers, web developers and enterprise architects to build data pipelines and model training infrastructure. You will also be responsible for publishing and presenting research to position Cambridge Assessment English as thought leaders in language assessment and learning, particularly in the application of artificial intelligence to these areas.


Identifying specific problems for junior machine learning researchers and data scientists to solve, you will also be a mentor to support their development as independent engineers.


With an eye on the bigger picture you will be engaging with policy debates about the governance and regulation of artificial intelligence technologies in educational contexts and responding to queries regarding the effectiveness of machine learning-based systems in specific contexts, and where needed, developing immediately deployable solutions as fixes.


What we’re looking for

The Machine Learning Engineer will work closely with the ELiT team and with other experts in machine learning at the Speech Research Group and the Institute for Automated Language Teaching and Assessment (ALTA). Our successful candidate will have excellent technical skills and knowledge related to speech processing, as well as being able to use these creatively and independently, to solve novel problems.


You will have the skills to help deploy AI-powered solutions widely in Cambridge Assessment English and Cambridge University Press products. To achieve this, you’ll articulate the benefits of applying machine learning to our contexts, and be able to communicate our vision to a range of stakeholders.



A postgraduate degree (e.g. MPhil) with a strong emphasis on machine learning or natural language processing (e.g. Advanced Computer Science; Machine Learning; Speech Technology; Informatics; Computational Linguistics) is essential. 

A doctoral degree (e.g. PhD) focused on research in Spoken Language Processing is desirable.



  • Expert computer programming skills.
  • Highly proficient at coding in Python and Java.
  • Communication skills to explain research and ideas to both specialist and non- specialist audiences.
  • Ability to work independently on technical and complex problems.
  • Ability to work with businesses to understand commercial goals and create solutions that use artificial intelligence.
  • Ability to develop approaches for deployable systems taking into account live deployment concerns such as computational complexity, latency, etc.
  • Ability to use, extend and maintain Linux OS scripts.
  • Strong understanding of speech processing techniques, such as speech segmentation, language modelling and automated speech recognition.
  • The ability, and willingness, to engage with existing software infrastructure.
  • Knowledge of social science research, particularly applied linguistics, is desirable.
  • The ability and willingness to lead multidisciplinary teams is desirable.



  • Conducting research on deep learning for sequence data, preferably including the modification of existing toolkits.
  • Using relevant deep learning toolkits (e.g. PyTorch; TensorFlow) to independently solve complex problems.
  • Designing statistical experiments and analysing the performance of machine learning models.
  • Publishing work in academic conferences/journals or industry circles.
  • Experience using language modelling toolkits (e.g. SRILM; KenLM) is desirable.
  • Experience using and creating dockers is desirable.
  • Experience using AWS services is desirable.
  • Experience of building speech recognition systems, preferably using Kaldi toolkit, is desirable.
  • Research experience in one or more of the following areas is desirable: spoken language assessment and feedback; large vocabulary speech recognition; spoken language processing (e.g. dialogue systems; summarisation); natural language processing (e.g. tagging; classification; sentiment analysis); speech synthesis (statistical or waveform based); audio processing.



  • Innovative – we are looking for someone who can learn complex techniques quickly and deploy them in creative ways, to meet changing needs in teaching and assessment around the world. You must be able to demonstrate how you have innovated and experimented with different approaches in the past, preferably in the application of ML.
  • Responsible – the products and services we build impact on millions of English learners. We are education experts trusted to uphold standards of transparency, fairness and quality. Our partners rely on us for reliable and robust systems; we expect our staff not to compromise on these standards.
  • Collaborative – the successful applicant must be able to collaborate effectively with small agile teams and larger organisations, including colleagues from different functional areas and backgrounds, such as academia, business development and software engineering.


The closing date for receipt of applications is 1 December 2020 and interviews will be held on week commencing 14 December


To apply please upload your covering letter and CV.


As part of your recruitment process you will be required to undergo a background screening. This will be carried out by our selected supplier, Credence Background Screening Limited.



Job Segment: Training, Process Engineer, Business Development, Developer, Education, Engineering, Sales, Technology, Research