ML / AI Engineer, Sentiment & Narrative Analysis

Role overview Build and operate sentiment and narrative classification models for Eastern European languages, where language and regional context can challenge tools designed for English. You will take models from research and evaluation into production, and report their performance with reproducible evidence.

Responsibilities - Design, train, and evaluate sentiment and narrative classification models for regional languages. - Measure model accuracy and prepare benchmarks comparing performance across tools. - Work with backend engineers to deploy models as production services. - Analyze errors by language, dialect, and subject area to find weaknesses. - Assess language models and fine-tune them when results support doing so. - Maintain versioned training pipelines and clear evaluation reports.

Requirements - 2–3 years of experience deploying NLP or sentiment models, with measurable results. - Strong Python skills and experience with PyTorch or TensorFlow. - Practical experience with transformer models and fine-tuning. - Knowledge of evaluation methods such as precision, recall, confusion matrices, and annotator agreement. - Familiarity with Eastern European, low-resource, or morphologically complex languages. - Clear, evidence-based communication.

Nice to have - Experience with multilingual or cross-lingual NLP, or work in social listening, media monitoring, or public affairs analytics. - Familiarity with MLOps tools such as MLflow or Vertex AI on GCP. - NLP research publications or open-source contributions.

Benefits and work setup Full-time, remote-first role with a regional-market competitive salary. The work is research-adjacent and may produce publishable, citable results.