Primary Responsibilities:
Support Ownership: Own the support function for our data science and bioinformatics algorithms in production, and make sure its answers are right and arrive when they are still useful.
Investigation and Root Cause: Lead complex investigations that cross a system boundary. Distinguish a biological signal problem from an algorithm problem, a data problem, and an infrastructure problem, and know which group owns the fix. Turn one-off investigations into durable rules and design changes, so the same question does not arrive twice.
Change Validation: Own process validation for pipelines and algorithms as operations change. Define what evidence is required before a change goes into a workflow that is running samples, and hold that bar when there is schedule pressure against it.
Team Leadership and Development: Lead, hire, and grow a team of Data and Bioinformatics Scientists. Lead by example, showing your team how to build and exercise deep systems judgment.
Cross-functional Partnership: Partner with Production Engineering, Product, Bio Development, laboratory operations, and the Data Science R&D teams to resolve operational issues and to prevent the recurring ones. Represent what the algorithms actually do, to groups that depend on them, and negotiate the commitments and the upstream fixes that need their agreement, not yours.
Ways of Working: AI is a routine part of the work here. Support is one of the places where these tools help most and where an unreviewed conclusion does the most damage, so you set the standard: what an agent may conclude on its own, and what has to be reproduced by hand before it goes in a record. We do not screen for prior experience with these tools, and many strong candidates come from environments where they were restricted; we provide the tooling and the ramp time.
What success looks like after a year:
•
The recurring escalations are fewer and test TAT reduced, and you can name which recurring issues you eliminated and how, including the upstream fixes and feature requests that made them stop arriving.
•
A service commitment exists for the function, agreed with the groups that depend on it, and the team is meeting it.
•
Change validation is a defined process the team follows, not a judgment someone makes fresh each time.
•
Your scientists handle investigations independently that used to come to you.
•
The production organization comes to your team for the answer, not for the route to whoever has it, and the team can settle an inconclusive report without handing it to another group.
•
Agent-assisted investigation is normal on the team, with a standard for it that you set.