Research Engineer

The role

You will be one of the first members of a small research team working close to data and the product. You will build evaluations from real data, run experiments that ship to production, and decide between building in-house or partnering with vendors.

What you will do

• Mine production workspace data into benchmarks and labeled datasets.
• Design label schemas and run labeling with inter-rater rigor.
• Build harnesses to capture expert judgment.
• Run experiments on offline workspace data and production systems.
• Evaluate frontier vendors and research teams.
• Publish research as the team makes meaningful progress.

Required qualifications

• Strong applied research background with depth in model evaluation, benchmarking, and/or failure analysis.
• Work samples or code demonstrating skills in eval frameworks, benchmark suites, failure-analysis reports, or tooling.
• Strong technical communication skills.
• Comfort with incomplete and ambiguous real workspace data.

Nice to have

• familiarity with simulation (Park et al.)
• human-in-the-loop evaluation (Scale HIL leaderboard)
• Cartridges and related context/memory-compression work
• memory for multi-party long-running settings
• agent observability standards (setting up Langsmith or similar)

Benefits

• Free Equinox membership & other health perks
• Generous equity grant vested over 4 years
• Health, dental, vision insurance

Before you apply

• 30 min intro call
• Technical conversation - walk us through an app you've shipped
• Paid take-home or IRL work trial

About the company

Ando is a messaging platform where AI agents take on work alongside their human teammates. They are rebuilding Slack around durable memory and agents as first-class participants.