11 hours ago
London, United KingdomStaff+
Responsibilities
- Design and ship real-time deep learning models for financial crime detection, including sequence-based, graph-based, and attention-based models.
- Define architecture strategy for model families, training paradigms, serving patterns, latency, and throughput optimization.
- Build reusable end-to-end pipelines spanning experimentation, training, and production deployment.
- Evaluate and prototype foundation-model and embedding approaches for transaction representation.
- Partner with Data Science on model evaluation, experimentation design, and causal measurement.
- Mentor engineers and data scientists on ML fundamentals, production practices, and architectural decision-making.
Requirements
- Production experience shipping deep learning models at scale under real-traffic and latency constraints.
- Ability to independently make and communicate architecture-level decisions involving model selection, training infrastructure, and serving strategy.
- Experience designing ML systems with hard latency and throughput requirements, including quantization, pre-computed embeddings, and batching strategies.
- Strong fundamentals in gradient dynamics, attention mechanisms, graph message-passing, and sequence modelling.
- Track record of influencing technical strategy across teams.
- Experience with Python, PyTorch or an equivalent framework, distributed training, and ML pipeline orchestration.
- Preferred experience in financial crime, fraud detection, AML, or regulated financial services.
- Preferred production experience with graph-based methods such as GNNs, entity resolution, and link analysis.
- Preferred foundation-model fine-tuning or LLM evaluation experience.
- Preferred experience establishing modern ML practices in organizations scaling their ML capabilities.
Benefits
- Starting salary of £145,000–£182,000 plus RSUs.
- Wise benefits package.
- Hybrid work arrangement indicated by the #LI-Hybrid designation.
- International, diverse, equitable, and inclusive work environment.
Categories
About Wise
Wise builds a cross-border payments and multi-currency account used by individuals and businesses to send, receive, hold, and spend money in 40+ currencies. Revenue comes from transfer and account fees, card interchange, and enterprise partnerships via Wise Platform, which lets banks and large companies embed its international payments. Founded in 2011 as TransferWise, Wise is a London-headquartered public company.