Data Scientist (AI/ML)
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Описание вакансии
TL;DR
Data Scientist (AI/ML/NLP): Lead technical development and optimization of adverse media screening services leveraging advanced data science, AI, and machine learning techniques with an accent on multi-language NLP, entity resolution, and regulatory compliance. Focus on developing and fine-tuning ML models, streamlining data pipelines, and building resilient real-time ML infrastructure.
Location: HK
Company
Leading global assurance partner providing assessment, advisory, inspection, and cybersecurity services to help clients manage risk.
What you will do
- Spearhead the technical roadmap for adverse media screening solution, implementing cutting-edge algorithms.
- Develop and fine-tune ML models and NLP systems for detecting adverse media across global multi-language sources.
- Streamline data processing pipelines for high-volume media sources.
- Collaborate with compliance, legal, and risk teams to meet regulatory requirements.
- Reduce false positives using advanced entity resolution and sentiment analysis.
- Architect resilient ML infrastructure for real-time monitoring.
- Champion testing, model validation, and coach technical teams.
Requirements
- Solid experience in applied data science, ML, and NLP, preferably in compliance or risk screening.
- Hands-on with adverse media analysis challenges like entity disambiguation and multi-language text processing.
- Proficiency in sentiment analysis, NER, classification, anomaly detection.
- Strong Python skills and experience with TensorFlow, PyTorch, Scikit-learn.
- Knowledge of big data technologies and near real-time systems.
- Ability to build explainable AI for regulated industries.
Nice to have
- Experience leading technical initiatives or managing teams in product environments.
Будьте осторожны: если работодатель просит войти в их систему, используя iCloud/Google, прислать код/пароль, запустить код/ПО, не делайте этого - это мошенники. Обязательно жмите "Пожаловаться" или пишите в поддержку. Подробнее в гайде →