Sardine is hiring a remote Senior Applied AI Research Scientist!
Sardine, which builds an agentic platform for fraud and financial-crime prevention, is hiring a senior applied AI research scientist to develop foundation models over its behavioral and transactional data. The company's datasets include device intelligence, behavioral biometrics, session telemetry, payment events, and consortium signals gathered from a large network of fintechs and banks. The person in this role will set the research agenda for next-generation fraud models, combining self-supervised pretraining and adaptation techniques with non-text sequential data. The position blends deep research work with production ownership: experiments must be evaluated honestly against strong classical baselines, then carried through pretraining, fine-tuning, distillation, quantization, and deployment on a low-latency real-time inference path. The scientist will also collaborate closely with engineering on training and serving infrastructure, with customer-facing teams to translate model behavior into practical risk decisions, and with legal, compliance, and customer model-risk groups to produce the explainability and documentation that regulated institutions require. The team is remote-first, with hires based in the United States or Canada, and compensation is set according to location.
Key responsibilities
- Identify high-value research opportunities, design rigorous experiments, and drive the roadmap for foundation-model development in fraud detection.
- Define and maintain evaluation standards, including offline benchmarks, time- and entity-aware holdouts, calibration, drift monitoring, and fair comparisons with classical baselines.
- Carry models end to end: data preparation, tokenization, pretraining, fine-tuning, distillation, quantization, and deployment in a real-time serving environment with tight latency constraints.
- Work with engineering on training infrastructure, GPU efficiency, feature and embedding stores, and production-scale model serving.
- Partner with customer-facing teams and clients to convert model capabilities and limitations into actionable risk decisions.
- Collaborate with legal, compliance, and customer model-risk teams on explainability, documentation, and governance for regulated bank and fintech customers.
Key requirements
- 4 or more years in applied machine learning, quantitative modeling, or ML engineering, including at least one foundation model that was pretrained or substantially adapted and deployed to real traffic.
- Hands-on experience with self-supervised pretraining, plus practical fine-tuning and adaptation.
- Production experience with model serving, versioning, monitoring, and rollback.
- Ability to self-manage ambiguous applied research projects and communicate clearly with data science, engineering, product, marketing, and external partners.
- Strong Python and SQL skills, with comfort preparing very large datasets.
- Nice to have: background in fraud, AML, payments, credit, or adversarial machine learning; experience evaluating LLM-based agents in production; publications, released models, or open-source work in representation learning or sequence modeling; familiarity with model risk management in a regulated financial setting.
- Seniority: Senior
- Work mode: Remote
- Key skills: Python, SQL, Deep learning, Foundation models, Self-supervised pretraining, Fine-tuning, Distillation, Quantization, Model serving, Model monitoring, Large-scale data preparation, Fraud detection, AML, Adversarial machine learning
Want more jobs?
Join our mailing list to receive updates about new data science and Senior Applied AI Research Scientist jobs.
