Full-Stack Software Engineer

Overview

Build RL environment tasks, graders and verifiers, and the data marketplace and customer sample tasks for an applied AI research lab that supplies training data and evaluations to frontier AI labs.

What you'll do

• Create many tasks in RL environments: start from an existing benchmark, work out which tasks the available codebases support, and keep the tasks consistent with it.
• Write two kinds of task: language-driven tasks (for example on legacy languages such as COBOL and Fortran), and outcome-driven tasks that state the goal without describing the code.
• Build the graders, tests and verifiers that score each task and make them hard to shortcut.
• Package tasks in Docker, in a standard format such as Harbor, and check them against agent runs.
• Build the data marketplace and customer sample tasks, and own delivery end to end.

What you'll need

• You have personally created tasks inside an RL environment, and can show a runnable repo or inspectable artifact and explain your part.
• Strong frontend skills.
• Strong backend skills.
• Linux.
• Docker.
• A testing mindset.
• Good judgment with AI coding tools.

Details

• Location: Mumbai, India (onsite).
• Compensation: ₹12-15 LPA + ESOPs.