Data Scientist II

About the role

See yourself at Radian? We see you here too. At Radian, we see you. For the person you are and the potential you hold. That's why we've embraced a new way of working that lets our people across the country be themselves, be their best and be their boldest. Because when each of us is truly seen, each of us gives our best - and at Radian, we'll give you our best right back. See Yourself as a Data Scientist II The Data Scientist II role sits on a team where data science products are core to what we build. We develop computer vision systems that analyze real estate properties, valuation models that price homes, generative AI that powers smarter property search experiences, and traditional machine learning that drives business decisions. We are also investing in systems that can reason, plan, and operate with increasing autonomy. This is a mid-level, hands-on individual contributor role for someone who wants end-to-end ownership. We're looking for someone who is a critical thinker and can work independently solving ambiguous problems with sound judgment and minimal direction, while remaining highly collaborative with teammates. You take full ownership of your work, demonstrating accountability from start to finish, and are self-motivated in driving projects forward without constant oversight. You communicate clearly, listen actively, and remain open to feedback and continuous growth. As part of a small, agile team, you will contribute end-to-end, rolling up your sleeves to execute rather than operate at a purely strategic level. You are action-oriented, enjoy solving problems, and bring a positive, engaging presence to the workplace-valuing both strong relationships and a sense of fun at work. Candidates should be prepared to share examples of production-grade models or systems they have owned end-to-end, including what they learned from deployment, monitoring, and iteration. Real estate or mortgage experience is not required; curiosity about how people search for, buy, finance, and value homes is helpful. See the Primary Duties and Responsibilities Analyze data to support (or disprove) a thesis - You'll dig into data, form hypotheses, and let evidence guide your conclusions. We value intellectual honesty over confirmation bias. Select and implement the right tools for the job - Not every problem needs a transformer. Some problems just need a well-tuned gradient boosting model. You'll know the difference. Build, train, test, and validate models - From algorithm selection to hyperparameter tuning to rigorous evaluation. You'll need solid grounding in math and statistics to evaluate model performance and defend your choices. Engineer models into production - This isn't research for research's sake. Your models need to run reliably in the real world, on real infrastructure, serving real customers. Document your work - Future you (and your teammates) will thank you. We maintain clear documentation for models, testing protocols, and decision rationale. Monitor and improve models in production - Models drift. Data changes. You'll keep watch and know when it's time to retrain, rebuild, or rethink. Explore agentic and reasoning systems - We're investing in semi-autonomous systems that can plan and act. You'll help us figure out what's hype and what's actually useful. Perform other duties as assigned or apparent. See the Job Specifications Basic Education and Prior Work-Related Experience: Degree Requirement: Bachelor's Degree or equivalent experience Work Experience: 2 or more years of prior work-related experience Primary Required Qualifications 2-5+ years of hands-on AI experience including working with LLMs (GPT, Claude, Qwen, or similar) via API/SDK and building and deploying ML or DL models in production environments. Core to success in this role is the ability to evaluate model performance beyond surface metrics and explain uncertainty clearly. This requires a strong scientific foundation in linear algebra, calculus, probability, and statistical inference. Understanding of prompt engineering, RAG architectures, fine-tuning approaches, and embedding models Strong command of supervised and unsupervised learning techniques: regression, classification, clustering, dimensionality reduction, ensemble methods Ability to evaluate LLM outputs critically and design appropriate guardrail systems. Familiarity with tokenization, context windows, and inference optimization Deep learning expertise including CNNs, RNNs/LSTMs, transformers, and attention mechanisms Practical experience implementing Reinforcement Learning algorithms: Q-learning, policy gradients, actor-critic methods, or multi-armed bandits. Understanding of reward shaping, exploration vs. exploitation tradeoffs, and temporal difference learning Ability to evaluate and define the appropriate model for each problem based on business requirements. Experience with model testing frameworks, model evaluation, validation strategies, and model documentation Other Required

Your Match

How well this role fits your profile.

Company Intel

What employees say

Worked at Radian Group? Share your experience