About The Opportunity
Building ML systems for compliance at a crypto exchange is a different kind of problem from most ML engineering work. The data spans on-chain transactions, fiat flows, KYC records, and behavioural signals that very few organisations have in one place. The problems are genuinely unsolved, the stakes are high, and the work has direct bearing on how a global exchange detects and responds to financial crime. For someone who wants their engineering work to matter beyond model accuracy metrics, this is an interesting place to be.
This role sits within a team of data scientists, analytics engineers, and compliance specialists who are building the analytical and AI infrastructure that powers the compliance function. You will work across the full ML lifecycle, from feature pipelines and model development through to deployment and monitoring, with close involvement from the domain experts who understand what the models need to do in practice.
AI-assisted development is how this team works. LLM-assisted coding, automated analytical pipelines, and AI-powered investigation tooling are part of the daily workflow. We are looking for engineers who already operate this way and who can raise the bar for what that looks like in a production compliance environment.
What You’ll Be Doing
- Design, build, and deploy ML models for compliance use cases including AML transaction monitoring, customer risk rating, KYC/KYB risk scoring, sanctions exposure detection, and SAR analytics, working closely with data scientists on model architecture and with data engineers on pipeline design.
- Own the production infrastructure for compliance ML: feature pipelines, model serving, monitoring, drift detection, and retraining workflows. Models in a compliance context need to be reliable, auditable, and well documented, and you will be responsible for making sure they are.
- Build and maintain internal ML tooling that the broader team depends on: reusable pipeline components, experiment tracking, model registries, and evaluation frameworks that raise the quality and speed of model development across the team.
- Apply AI-assisted coding and automation as a matter of course: using LLM tooling to accelerate development, building automated pipelines that reduce manual analytical work, and integrating LLM-based capabilities into compliance workflows such as SAR narrative assistance, alert summarisation, and investigative triage. The expectation is that you bring this fluency with you, not that you develop it here.
- Work with data scientists to take research-stage models into production, reviewing feature logic, validating pipeline assumptions, and bridging the gap between a notebook and a deployment that a compliance team can rely on.
- Collaborate with compliance and legal stakeholders to ensure models are explainable and documented to the standard required for internal governance and regulatory review.
- Keep a close eye on developments in compliance-relevant ML: graph neural networks for network-based AML detection, anomaly detection approaches for novel typologies, and emerging LLM applications in regulated environments, bringing relevant ideas into the team's work where they hold up to scrutiny.
What We Look For In You
- 8+ years in ML engineering, data science, or a closely related field, with a strong track record of taking models from prototype to production in environments where reliability and auditability matter. We welcome candidates across seniority levels; scope will be calibrated to your experience.
- Solid Python and experience with ML frameworks alongside hands-on MLOps practice including model deployment, monitoring, and CI/CD pipelines for ML workflows.
- Experience with big data platforms such as Spark, Databricks, Hadoop, or MaxCompute, and comfort designing and optimising both batch and real-time data pipelines that feed production ML systems.
- Demonstrable fluency with AI-assisted development: you use LLM coding tools before, you have built LLM-integrated pipelines or automation workflows in a professional context, and you have a clear, practical view on where these tools genuinely improve engineering output and where they introduce risk.
- A good working knowledge of ML fundamentals across supervised, unsupervised, and anomaly detection methods, and an understanding of how different approaches translate to compliance problem formulations.
- Familiarity with explainability frameworks such as SHAP or LIME, and an appreciation for what model governance and auditability require in a regulated environment.
- Good communication skills and a collaborative working style, with the ability to work effectively alongside data scientists, compliance domain experts, and engineers who each bring a different perspective to the same problem.
- Experience in financial services, fintech, or a crypto exchange, particularly in AML, KYC/KYB, transaction monitoring, or a related compliance domain, is a meaningful advantage.
- Familiarity with the crypto ecosystem, on-chain data, blockchain analytics, or VASP regulatory frameworks is a plus and will give you a head start in understanding the data you will be working with.