Staff Applied Machine Learning Engineer, Financial Crime — Hybrid, London, United Kingdom

Overview

Set technical direction for machine learning models that detect financial crime in cross-border payment activity. This staff-level role focuses on production risk modeling and the architecture needed to apply modern learning methods reliably. You will work with data scientists, platform engineers and domain specialists from research through deployment and impact assessment.

Requirements

• Candidates need experience deploying deep learning systems to real traffic under demanding latency and throughput constraints.
• Independent architectural judgment, strong learning fundamentals and a record of influencing technical strategy are required.
• Python, a framework such as PyTorch, distributed training and pipeline orchestration are essential.
• Financial crime, graph methods and foundation-model experience are advantages.

Responsibilities

• Design and release deep learning models for real-time decisions, including sequence, graph and attention-based approaches.
• Choose appropriate training and serving patterns and explain the tradeoffs behind them.
• Build reusable pipeline conventions so future models can move from experimentation into operation with consistent engineering discipline.
• Explore transaction representations using embeddings and foundation models.
• Work with data science colleagues on evaluation, experimental design and measurement where conventional A/B testing is difficult.
• Mentor engineers and researchers in production practices and architectural reasoning, helping the wider financial crime organization develop a stronger modeling foundation.

Conditions

• This full-time role is hybrid in London.

Tech stack

• Python
• PyTorch or equivalent
• Distributed training
• ML pipeline orchestration