Research Project on Context-adapting Semantic Learning for Large Speech and Language Models

Fecha: 16 jun. 2024

Ubicación: BARCELONA, ES

Empresa: Telefonica S.A.


Telefonica Innovación Digital - Discovery Research is hiring for a research project focused on developing and integrating applications related to context learning for large speech and language models .

We are looking for a Research Engineer to join our team of experts who are working in the development of machine learning solutions to contribute on enhancing user experiences. During this project the candidate will be responsible for implementing and deploying context-adapting learning solutions by using deep learning-based speech and language models. The machine learning techniques employed will help in building powerful and adaptive speech and language models, capable of handling complex linguistic phenomena in unstructured dialogues. These solutions will also aid in the development of advanced voice assistants systems, capable of recognizing and aligning to users’ context and preferences, in a wide range of environments and with high accuracy. Overall, the goal is to create highly scalable and robust speech/language processing systems, capable of delivering highly personalized, trustworthy and contextually relevant results for various applications, including customer service, virtual voice assistants, and other related domains. 


Telefonica Innovación Digital - Discovery Research is a leading industrial research lab based in Spain. At Telefonica Research group, we follow an open research model in collaboration with Universities and other research institutions, promoting the dissemination of scientific results both through publications in top-tier peer-reviewed international journals and conferences and technology transfer. Our multi-disciplinary and international team counts several full-time researchers and engineers, holding MSc. and PhD degrees in various disciplines of computer science and electrical engineering.  


Key Responsabilities: 

  • Develop and deploy personalized and adapted-to-context solutions for keyword spotting, automatic speech recognition and large language models, based on deep learning techniques.  

  • Implement state-of-the-art methods such as semantic knowledge graphs, privacy preserving learning and early-exit architectures to ensure context adaptation, human-preferences alignment, data privacy and federation of heteregeneous speech/language deep learning architectures. 

  • Conduct experiments and evaluations to assess the performance of solutions and optimize algorithms for deployment in low resourced environments. 

  • Stay up-to-date with industry trends and developments related to computational linguistics, speech and language models and propose recommendations for future improvements. 

  • Collaborate with cross-functional international teams to integrate speech and language solutions into existing platforms and applications.  

  • Write technical reports and documentation to communicate results to internal and external stakeholders. 



  • MSc. in Computer Science, Electrical Engineering, or related field. 

  • Strong understanding of digital signal processing and machine learning algorithms and their implementation, like as short-time and fast-fourier transforms and encoder-decoder or transformer architectures. 

  • Strong programming skills in languages such as Python and Pytorch. 

  • Strong communication and interpersonal skills to collaborate with cross-functional teams. 


If you join us, we offer you...

  • Work in a dynamic team and be part of challenging projects that will allow you to play a leading role in the transformation of the company.
  • An interesting package of social benefits (luncheon vouchers, health, life and accident insurance, pension plan, flexible compensation plan...).
  •  A flexible working environment with hybrid models of face-to-face and telecommuting.
  •  Access to a wide range of continuous learning opportunities for you to grow with us.
  • Participate in volunteering programmes, social actions, international mobility, off-site...



If you join Telefónica
You join almost 100 years of history, a team of 106 nationalities present in more than 35 countries. You join a team that works to connect people wherever they are, without borders. A team that is leading the digital revolution with the enthusiasm of the first day in all our businesses, creating the best digital ecosystem for our clients: Network, IoT, Cloud, Cybersecurity, Innovation, etc. At Telefónica you have everything you need to create the best version of yourself. We need people like you to join this great challenge, who want to create the Telefónica of tomorrow.
At Telefónica we are committed to new ways of working and we are leaders in the implementation of Digital Disconnection under the "Disconnect to Reconnect" principle.
You join a company whose activity is governed by its code of ethics, Our Responsible Business Principles. We are looking for people who identify with them, who help us make decisions based on integrity, commitment and transparency and who are committed to ethical management, promoting fairer and more sustainable social and environmental development.
We strongly believe diverse and inclusive teams become more innovative and they achieve improved results. Telefonica aims to promote and guarantee a place for everyone, just the way we are: gender, age range, sexual orientation and/or identity, culture, creed, disability or any other personal condition.