2 min lesson
The charter in one frame
In "The charter in one frame", respond to "What is the primary success metric for a Cursor Research Scientist?" Give the answer first and the evidence second.
Step 1 of 2
- Why the role exists
- Make the next shipped coding-agent model better via RL on real user data
- Primary lever
- Reinforcement learning at scale on the ComposerCursor's own fast coding model, tuned for the editor and priced well below frontier models; the recommended day-to-day model for executing a plan. Press Enter for the full definition. family and Cursor TabCursor's original autocomplete: multi-line, edit-aware suggestions you accept with the Tab key. Press Enter for the full definition.
- Success metric
- Production model improvement that reaches users - not a paper or an isolated benchmark number
- Where it sits
- SF and NY, full-time, inside the Engineering org
The org placement is a tell. This role lives under Engineering, not in a separate research division that hands findings over a wall. Research and engineering are not different castes at Cursor: you are expected to build the training, eval and data systems your hypotheses need, then ship the win yourself.
In a paper-first lab, a successful quarter can end in a submission and a strong eval table. Here the loop from idea to shipped model is short and the result is judged by whether it made the deployed model better for millions of developers. If you internalize one thing about this charter, make it that the deliverable is a better model in users' hands and the eval number is only useful insofar as it predicts that.
“I read this role as: own RL research for Cursor's coding agents end-to-end and be measured by whether the next ComposerCursor's own fast coding model, tuned for the editor and priced well below frontier models; the recommended day-to-day model for executing a plan. Press Enter for the full definition. or Tab model is actually better in production - not by what I could publish about it.”