NLP Specialist

  • Montreal / Quebec City / Sherbrooke
  • Research and Development
  • Full-time
See All Positions

Most websites out there are trying to give out the most information at once, sometimes to impress, but sometimes also to cover the fact that they don't know what you might be looking for in the first place. Some giants in the Commerce and Service fields have raised the bar of customer expectations really high, and many retailers created their pages trying to reach that standard, often to no avail. From the moment you buy an item online, and more importantly, when you need support with the said item you've bought, most of the data you leave behind isn't used to leverage a better customer service. The experience isn't guided, or optimized; you're simply floating through without a rudder, having to start with new requests or queries each time, even if you are regular customers of the website, and barely leveraging what could make a good experience a great one.


Become the navigator of these customers' journeys.


The Coveo platform is a great tool to help businesses leverage the data that customers create using their commerce website and reaching out to their customer service's solutions. Not only are we helping both these services in their own rights with state-of-the-art Natural Language Processing (NLP), recommandations and text classification, but we bridge the two worlds by using a common database to be able to track a customer from the moment they are shopping, buying, troubleshooting and replacing a product. And that's why we need NLP Specialists. Researching and building NLP models to help understand customer’s textual inputs is crucial to our product and to recommend the right solution to their problem. So if tweaking with the BERT architecture is something you are familiar with, that you want to tackle some major natural language understanding (NLU) and natural language generation (NLG)  problems, we want to talk with you!


What's a normal day like?


  • During your morning scrum, your team gets to talk with the Product Manager about some specific needs of a new client. You throw in a few ideas on how we could actually make it happen.
  • As you read an article on the latest language model that a colleague sent your way for a project you have coming up, one of the back-end developers working on the deployment pipeline of your projects sends you a high five on Slack, as you gave them a great help doing a part of their job yourself, as they were in a tight crunch.
  • You tweak a bit on a model that you've made for a client: it was brought to your attention that your training dataset of that model might have contained a blind spot that doesn't consider a part of new products for the recommendations made. You also take some notes for the post-mortem that will be organized to review what could have been done better and solve the root cause of that issue.
  • As you take a break for lunch, you watch a webinar that one of your colleagues made for the ACL Conference. Seeing her present right after Alibaba's data science team's presentation, and having as many viewers as them really hits a proud spot
  • Your afternoon revolves around this new project your team has: the answer extraction systems we have running are great, but a bit limited as to the format they can search through. As most Knowledge management databases of our Customer Service solution include various other types of files, it becomes a mighty challenge that has to be addressed.
  • Right before you call it day, you receive a feedback from the Customer Success team: The project you've spent the last quarter working on

Directly from your research to application


  • Is this the type of role where you will redact a lot of research paper and become an academic staple? Not really.
  • Will you have to present the result of your work in front of people in big conferences? Perhaps a few times, if you want to.
  • Will your work and learning be directly tested on the field on real use cases and databases for people that will directly benefit from your efforts? Definitely
  • Will there be some mistakes, trials and errors? Sure: it's R&D work after all! We actually expect to see some, otherwise you're not trying hard enough!

At Coveo, we believe in tackling difficult problems. While some say "It's too hard, it's impossible", we're saying "Challenge accepted, we will find a way". And we can test our solutions first-hand. That's why we need a team player that will do what is needed to get the project up and running in production.


What do we expect for the role?


  • A strong knowledge base backed by a Masters or a PhD
  • At least 2 years of experience in the industry
  • Familiarity with state of the art deep learning architectures: Recurrent Neural Network (RNN), Multilayer Perceptron (MLP), Convolutional Neural Network (CNN), Transformers.Familiarity with the state-of-the-art NLP models: the BERT family (BERT, DistilBERT, ALBERT, RoBERTa), ELECTRA, GPT.
  • Proficiency with either PyTorch or TensorFlow, and NLTK or spaCy are a must. HuggingFace would be an asset too.
  • Proficiency in data analysis: using notebooks, PySpark, Pandas and navigating through a SQL request should be simple, for example.
  • Experience with text classification, semantic textual similarity or question answering.
  • Experience with multi-lingual models is a plus.
  • Go beyond the notebook and implement deployable NLP models in a production environment.

Do you think it’s a fit?  Send us your resume and convince us that you are a must-have rather than a nice-to-have.
Join the #CoveoLife!

What is Coveo?

Coveo makes business personal. Coveo uses AI technologies and intelligent search to personalize every digital experience for customers, partners, dealers, and employees. Coveo combines unified content, unified interactions behavioural data and machine learning to deliver relevant information and recommendations across every business interaction, making websites, e-commerce, contact centres and intranets efficient, effortless, content-rich, thus boosting conversion. Coveo is also embedded in many leading business applications from vendors including Microsoft Dynamics, Salesforce, ServiceNow, Sitecore, Xero and more.

What is it like to work at Coveo?

If it were easy, someone else would have done it. All of our colleagues at Coveo find the drive to continuously seek new challenges and test roads no one else has ever explored. This ethos has allowed us to become a world leader in an innovative industry and to create a collaborative, diverse and trusting place to work filled with amazingly talented and passionate people. This is also why Coveo was named Employer of the year AND Company of the year 2019 by the Quebec Chamber of Commerce!