Magnet.me  -  Het slimme netwerk waarop hbo‑ en wo‑studenten hun baan of stage vinden.

Het slimme netwerk waarop hbo‑ en wo‑studenten hun baan of stage vinden.

Deze vacature is verlopen. Je kunt daarom niet meer liken of solliciteren.

Vergelijkbare vacatures bekijken

NLP Data Scientist/Software Engineer

Baan Thuiswerken
Geplaatst 26 mrt. 2024
Werkervaring
0 tot 5 jaar
Full-time / part-time
Full-time
Functie
Salaris
€ 50.000 - € 65.000 per jaar
Soort opleiding
Taalvereiste
Engels (Vloeiend)

Je carrière begint op Magnet.me

Maak een profiel aan en ontvang slimme aanbevelingen op basis van je gelikete vacatures.

We are looking for a Data Scientist with Software Engineering skills, or a Software Engineer with Data Science skills, with a love for complex algorithms and Natural Language Processing. You will be working on the core of Deepdesk: the AI that produces live textual suggestions to helpdesk agents.

  • The most important techniques used are: clustering, classification and language models.
  • Also important are normalization/preprocessing, customer-controlled rewrite rules, benchmarking and reporting.
  • You will tweak and tune the logic and conduct experiments to improve the quality and speed of the suggestions.
  • You have the end-user in mind - what will help him or her to work faster and better?
  • You love algorithmic code. You will work together with the other data scientist(s) at Deepdesk, tackling problems together and offering support at the engineering side of things.
  • Affinity with MLOps, data engineering, Kubeflow is a pro.

Languages/technologies: Python, Tensorflow, Pytorch, Scikit-learn, Kubeflow, GCP,

ML techniques: Clustering, classification, language models (GPT-2, LSTMs, ngrams), word embeddings, TFIDF, NER/anonymization.

More on our Stack: https://stackshare.io/deepdesk/machine-learning

Everyone uses AI. All-day, every day.
At Deepdesk we use the power of NLP to empower human agents with AI.
We have successfully picked up the task to train the first Dutch GPT-2 base model from scratch. Since no Dutch base model existed, we did it ourselves.
Our technology is trained with 3+ million interactions and is actively used in 20.000 conversations every day.
We are innovative, driven, and ready to grow!

ICT
Amersfoort
Actief in 1 land
15 medewerkers
90% mannen - 10% vrouwen
Gemiddeld 38 jaar oud