Turn this role into an interview — a resume and cover letter built around what this employer wants.
Phantom ищет опытного руководителя ML-инженерии для определения долгосрочной дорожной карты Growth и Engagement ML систем. Вы будете проектировать и внедрять production‑grade пайплайны и решения в реальном времени для персонализации, уведомлений и онбординга.
Кандидат с глубоким знанием Python/Scala/Java и опытом в PyTorch/TensorFlow и распределённых систем, умеющий связывать улучшения алгоритмов с бизнес-метриками MAU/DAU, конверсией и удержанием, приветствуется.
Phantom connects people to open global markets, including perpetuals, prediction markets, tokenized assets, stablecoins, and memes. Its app provides access to real-time market data and verified trader performance, with self-custody and open networks at its core. Phantom also partners with companies in finance to make financial products accessible to everyone.
Define the long-term technical roadmap for Growth and Engagement ML systems, ensuring scalability, reliability, and measurable business impact Architect and deploy production-grade ML pipelines and real-time decisioning systems for personalization, notification dispatch, and onboarding flows Evaluate and integrate machine learning techniques, including multi-armed bandits, reinforcement learning, LLMs for content generation, and graph neural networks Design, train, and validate models for churn propensity, lifetime value forecasting, next-best-action, and lookalike modeling Build and optimize recommendation engines and semantic search systems to surface relevant content, products, or features Establish experimentation frameworks, including advanced A/B testing, causal inference, and multivariate testing, to validate model variants in production Partner with Product and Growth Marketing teams to translate business hypotheses into actionable machine learning problems Mentor and coach senior engineers across data and ML organizations Advocate for ML engineering best practices, including model monitoring, feature store utilization, reproducible training pipelines, and data governance
Будет плюсом: No additional preferred qualifications specified