Artificial Intelligence Engineer

NLP High Importance:

Model training. The candidate should have previous experience in training and evaluating transformer-based models (including more modern architectures like ModernBERT) for downstream tasks with low data volumes.

Data collection. Experience in conducting human annotation tasks. Preferably experience in synthetic-data collection for training models in downstream tasks.

Information Retrieval. Knowledge about information retrieval systems, including BM25, embeddings models and re-ranking models. Should know basic metrics for evaluating information retrieval like Recall, Precision, MRR and nDCG and where they are applicable.

RAG and LLMs. Experience with handling retrieval augmented generation tasks with third-party / open-source LLMs for generating a response. Knowledge in grounding and hallucination checks.

General. Knowledge of the following used throughout the project Topic modelling, clustering, text-classification, NER (preferably models like GliNER).Platforms / Tools High Importance:

AWS; OpenSearch; Lambda; ECS; CI/CD; Docker

FastAPI Torch; Spacy; TransformersFrontend Development Low to Moderate Importance:

Ideally someone who is full-stack, i.e. able to deploy and maintain a frontend according to business user demands for a chat application. Can continue along the same lines as Andreas solution which uses FastHTML OR alternatively for more performant frontend solution someone with actual professional experience in the matter.Soft Skills:

Stakeholder engagement High Importance

Previous experience in liaising with stakeholders to understand project needs. Pro active in such a way that they take the initiative to understand where business value lies in the process and how to translate that to technical requirements. Communication High Importance

Ability to communicate technical requirements to business and act as their tech point of contact when exploring new solutions or data sources.