Lead Data Scientist, Rail Data and Risk

Role Summary

As Lead Data Scientist, Rail Data and Risk, you will build Sperry's view of where risk sits in our customers' track and how it is changing. We run non-stop inspection across North America and hold years of ultrasonic, induction, and eddy current test data. Your first job is descriptive: where defects and surface conditions are concentrated, and how they are trending.

The second is predictive: where failure is likely, and what it would cost. These are different problems, and the expectation is that you compound toward all of them rather than arrive expert in each. What makes the role senior is the second half of it. The analysis only counts once it reaches the people who act on it - our commercial team, the analysts reviewing that track, and our customers.

This is the first seat in a new US data science team. You will hire and line manage a data scientist working on analyst decision quality and a software engineer who puts models into production. Both join in the following quarter, so for the first two quarters this is majority hands-on work, with the team building around you rather than you waiting for it.

The role carries real autonomy. The business case, the reporting line, and the recruiting channel are already in place; the methodology, the models, and the way this work reaches its audiences are yours to set. We hold a large amount of data and not everything we would ideally want, so part of the job is getting useful answers out of imperfect data and telling us what we should be collecting instead.

We expect an exceptional level of drive and ambition. You think beyond today's work to what the team and organization need next, champion bold ideas, and see them through. Your hunger is infectious – it inspires those around you to aim higher. We are looking for someone who sets the standard for humility. You are open and receptive to feedback and criticism, you create an environment where everyone feels valued and heard, and you handle both success and failure with grace.

We need someone who can operate with full autonomy. You will set your own direction aligned with organizational goals, define and deliver ambiguous, high-impact work, and be trusted for critical decisions. You should be capable of representing the team externally, including in front of customers. Strong analytical thinking is critical. You will decompose ambiguous, multi-faceted problems, reason about complex data flows, and anticipate the downstream consequences of methodological choices.

You translate business problems into defensible technical answers, and you are candid about what the data will and will not support. You will be building a team, not just doing the work. We expect you to hire well, coach, set targets, measure performance objectively, and develop the two people who will join under you. You should be able to question, challenge, and improve existing processes, and you should have the confidence and gravitas to do that with staff and customers at all levels.

• Trend Sperry's multi-modal test data over time - internal flaw detection, induction, and eddy current - for defect growth and surface condition degradation across our non-stop inspection programs in North America
• Connect raw test measurements and their metadata to the physical conditions they represent: internal defects, surface conditions, rail flaws
• Analyze defect and error types and frequency by subdivision to identify where risk is concentrated and how it moves
• Apply risk-based models to estimate the probability and consequence of failure
• Own the KPIs for the monthly operational review and customer account review meetings
• Present findings to the commercial team in a form they can use in account conversations
• Work with the analysis organization so that what the data shows about defect and surface-condition patterns reaches the analysts reviewing that track
• Present risk findings to customers alongside the commercial team
• Identify gaps in current data collection and recommend what Sperry should capture to support better analysis
• Validate findings against field conditions with track engineering and testing teams
• Hire and line manage the two further seats in the pod, and direct their work
• Write and maintain documentation so that the analysis is transferable rather than held tacitly

• Statistical depth: probability, hypothesis testing, regression analysis, time series analysis
• SQL and Python, or equivalent…