Data Scientist

📋 Description

• • Shape the future of fraud prevention: As a Data Scientist at GeoComply you will design, train, and deploy machine-learning models that protect more than 400 million devices and 12 billion annual transactions from sophisticated cyber-criminals. Your work directly determines whether a bettor in Las Vegas, a banker in London, or a gamer in Lagos is shielded from account takeovers, bonus abuse, location spoofing, and money-laundering schemes.
• • Own the full ML lifecycle: From raw telemetry to production inference you will architect end-to-end pipelines—feature engineering on petabyte-scale data, hyper-parameter tuning with Bayesian optimization, containerized model serving, and real-time A/B testing—ensuring every line of code you ship runs at sub-second latency and 99.99 % uptime.
• • Turn adversaries into algorithms: Collaborate with fraud analysts who reverse-engineer the latest botnets, VPN rings, and device-farm tactics; you will translate their findings into gradient-boosted ensembles, graph neural networks, and anomaly-detection transformers that evolve faster than the threats.
• • Optimize for precision and fairness: Balance razor-sharp precision-recall trade-offs while minimizing false positives that could lock legitimate users out of their accounts. You will build explainable-AI dashboards (SHAP, LIME) that give regulators clear evidence and give users transparent reasoning.
• • Build data foundations that scale: Construct scalable Spark and Databricks workflows that ingest 50 k events per second, cleanse geolocation signals, and generate hundreds of behavioral features—from gyroscope jitter to TLS fingerprint entropy—ready for both batch retraining and real-time scoring.
• • Champion MLOps best practices: Implement CI/CD pipelines for ML models, automate drift detection, and orchestrate blue-green deployments so new fraud signatures are blocked within minutes, not days. Your Git history and model-cards will become the gold standard for reproducible research across the company.
• • Collaborate across borders: Partner with engineers in Vancouver, product managers in London, and compliance officers in Sydney to translate regulatory requirements into technical specs. Your Slack threads and design docs will bridge time-zones and cultures, ensuring every stakeholder understands how your model reduces risk.
• • Drive continuous innovation: Allocate 20 % of your time to R&D—experiment with Large Language Models for synthetic fraud-data generation, test graph-based clustering to uncover mule networks, or prototype reinforcement-learning agents that adapt to zero-day exploits. Your white-papers and conference talks will elevate GeoComply’s thought-leadership in cybersecurity.
• • Mentor and be mentored: Lead knowledge-sharing sessions on advanced gradient boosting, coach junior data scientists through Kaggle-style hackathons, and receive executive-level mentorship on product strategy. Your growth plan includes an annual professional-development budget for courses, certifications, and global conferences.
• • Measure real-world impact: Track KPIs such as fraud-loss reduction, model-lift over rules-based systems, and customer NPS improvements. Present quarterly business reviews to the C-suite showing how your algorithms saved millions in chargebacks and regulatory fines.