Staff Data Scientist, Methodologies
Posted 4hrs ago
Employment Information
Report this job
Job expired or something wrong with this job?
Job Description
Staff Data Scientist developing foundational deep-learning methods for Tilt’s mobile-first credit products. Building models and specialized agents to improve lending decisions and financial outcomes.
Responsibilities:
- Develop and test modeling approaches using deep learning architectures such as transformers, embeddings, and neural networks
- Build and refine frameworks integrating with existing ML pipelines for credit modeling and loss prediction
- Collaborate with data scientists, engineers, and product partners to translate complex ideas into practical solutions
- Experiment with pretraining and fine-tuning representation learning techniques on financial data
- Share findings and mentor others to grow deep learning capability
- Measure model impact through customer and partner outcomes, not only accuracy
- Use agentic development tools such as Claude Code and Codex for experimentation, feature engineering, and model iteration
- Design and ship specialized agents supporting data science and business workflows
- Help advance foundational modeling approaches across Tilt’s products and improve access to fairer credit
Requirements:
- Experience with deep learning, including neural networks, embeddings, or transformers
- Comfort working with modern ML frameworks such as PyTorch or TensorFlow
- Experience applying machine learning to real-world problems, ideally in financial services
- Strength connecting data science to measurable business outcomes
- Clear, thoughtful communication of technical ideas
- Bonus: hands-on experience building with LLMs or agent frameworks, or excitement to learn
- Willingness to learn, experiment, and collaborate
Benefits:
- WFH office reimbursement
- Competitive compensation packages
- Generous equity
- Flexible health plans at every premium level
- Substantial health-plan subsidies
- Direct exposure to the leadership team
- Paid global onsites twice yearly
- Growth opportunities based on contributions rather than rigid promotion timelines


















