Talking the talk, walking the walk – The NLP journey of a major Swedish bank

At the Swedish bank SEB, we have experimented with NLP for quite some time. This talk describes our journey from simple statistical models, via language models to Large Language Models

• Limitations of ML when processing text.
• Difficulties around using sensitive data for NLP.
• How NLP has evolved at SEB, and some of our challenges.
• How we try to deliver business value with LLMs

————————————————————————————————————————————————————

Manne Fagerlind – Technical lead – Machine learning operations | SEB

Earlier in his career, Manne was a postgrad student in linguistics but moved on to become a backend developer and software architect. He has worked in a wide range of industries in these roles and spoken at many conferences, about topics like Java technologies, software craftmanship and MLOps. Six years ago he switched his professional focus to machine learning and big data and eventually came to the MLOps team at SEB. Working with NLP gives him the chance to combine two of his greatest interests: human language and machine learning.