Database Systems Researcher (Research Coding Tasks)

What You'll Do8

• Convert a recent paper with public code, ideally your own, into a self-contained coding task an AI agent can attempt.
• Remove the code that carries the paper's main contribution and package everything else as the starting scaffold.
• Write the task statement so an agent understands the goal without seeing the removed implementation.
• Define held-out evaluation workloads that the agent cannot inspect ahead of time.
• Set per-metric pass thresholds tied to the results reported in the paper.
• Build a hidden reference solution that restores the missing contribution.
• Provide a second working solution that clears the bar, proving the thresholds are reachable.
• Evaluate agent attempts on your task and judge whether each one meets the defined metrics.

Requirements7

• Published database systems research at venues such as SIGMOD, VLDB, ICDE or CIDR, or a comparable record.
• Public code released alongside at least one paper, with a GitHub repo or artifact link you can share.
• PhD in computer science or a related field, either finished or still in progress.
• First-author papers that report measured gains over a baseline stand out.
• Comfort with Docker and benchmark setups another researcher can rerun.
• Working depth in query optimization, storage engines, indexing or learned database components.
• Willingness to complete a background check as part of the application.

Who Should Apply

PhD-level database researchers and research engineers who publish at systems venues and release runnable code with their papers. You should be able to look at your own implementation and decide which piece an agent could realistically rebuild, then write the thresholds that prove it did.

Salary Insight