ML Engineer — LLMs
Il y a 5 heures
belgium belgium
Chut
Temps plein
Gratuit avec email ou Google
Enregistrez cette offre et organisez votre recherche
Créez un compte gratuit pour enregistrer des offres d'emploi, créer des alertes et revenir à cette liste depuis votre tableau de bord.
Gratuit avec email ou Google
En continuant, vous acceptez nos Conditions d’utilisation & Politique de confidentialité.
ML Engineer — LLMs 5+ years Full-time Belgium / remote-friendly You own the model layer: continued pretraining, fine-tuning, evaluation, retrieval, and serving — on infrastructure we control.
What you'll do
* Run continued pretraining (CPT) and supervised fine-tuning (SFT) to adapt LLMs to our domain, in Dutch, French and English.
* Build evaluation harnesses that tell us whether a model change actually helped before a customer sees it.
* Design retrieval and grounding so the agent answers from the customer’s real data, not from vibes.
* Optimise inference — quantisation, batching, serving — on our own GPU servers. What you bring
* 5+ years in machine learning engineering, with recent hands-on LLM work.
* Practical experience with CPT, SFT, evaluation methodology, and inference serving.
* Strong Python and comfort working close to the GPU.
* A healthy scepticism of benchmark numbers and a habit of checking real outputs.
* High capacity to work with AI tools — they are part of how we build, not a novelty. Languages: English. Apply — br
What you'll do
* Run continued pretraining (CPT) and supervised fine-tuning (SFT) to adapt LLMs to our domain, in Dutch, French and English.
* Build evaluation harnesses that tell us whether a model change actually helped before a customer sees it.
* Design retrieval and grounding so the agent answers from the customer’s real data, not from vibes.
* Optimise inference — quantisation, batching, serving — on our own GPU servers. What you bring
* 5+ years in machine learning engineering, with recent hands-on LLM work.
* Practical experience with CPT, SFT, evaluation methodology, and inference serving.
* Strong Python and comfort working close to the GPU.
* A healthy scepticism of benchmark numbers and a habit of checking real outputs.
* High capacity to work with AI tools — they are part of how we build, not a novelty. Languages: English. Apply — br