Description
As Harvey’s AI product changes quickly, teams need to distinguish genuine improvements from features that simply look impressive in a demo.
The Product Data Scientist helps create that evidence.
Rather than functioning as a reporting service, this position is embedded with Product, Engineering, Design, Go-to-Market, and company leadership. The goal is to define how Harvey measures user value, adoption, engagement, retention, and the impact of changes to models and workflows.
An important part of the role is establishing metrics that remain useful while the product itself continues evolving. The scientist will help define north-star measures and product-level indicators, build source-of-truth reporting, and prevent teams from optimizing around vanity metrics that do not reflect actual customer value.
Experimentation is another core area. Product changes may involve interface design, workflow automation, model behavior, agent capabilities, or different ways of presenting AI-generated results. The data scientist will design A/B tests and causal analyses to determine which changes actually improve outcomes.
Because Harvey is an AI product, model changes can create unusual analytical problems. A different model may affect accuracy, response behavior, latency, user trust, task completion, or downstream usage simultaneously. Measuring these trade-offs requires both statistical rigor and an understanding of how users interact with generative AI.
The position is particularly suited to someone who enjoys turning ambiguous product questions into measurable hypotheses. Strong SQL and analytical skills are important, as are experimentation, causal inference, metric design, and communication with non-data stakeholders.
Experience with AI/ML products, enterprise software, developer tools, or complex professional workflows is valuable but not mandatory. Harvey also values people who can operate when the underlying datasets and product definitions are still being built.
This is one of the more product-oriented AI opportunities in this batch: the job is less about training models and more about helping the organization understand whether its AI is creating durable value for real users.
Curated opportunity. Please verify details and apply via the original link below. No Signals are required for this project/job.
https://www.harvey.ai/company/careers/1321b20d-d6f8-4bdb-b2de-ec33c0948b53