Research Engineer - Code Generation & Model Evaluation

What you would do

• Analyze and resolve bugs across diverse codebases in Python, Java, Rust, C++, Go, and TypeScript
• Contribute design and implementation of new features and code generation workflows
• Refactor code to boost performance, readability, and maintainability for production systems
• Develop real-world coding tasks and review AI model solutions for accuracy and quality
• Author clear technical annotations and feedback to drive model training and improvement

Who they want

• Deep expertise in competitive programming and problem-solving across multiple languages
• Hands-on expertise in one or more languages (Python3, Rust, Go, Java, C++, TypeScript) combined with strong algorithms and data structures knowledge
• Proven open-source contributions or collaborative software projects
• Strong analytical ability to interpret complex constraints and reason about solution trade-offs
• Meticulous attention to code quality, validation, and output consistency in fast-paced environments

Main skills

What the interview asks about

1. 1.Debugging multi-language codebases

   AI code spans multiple stacks; you must isolate problems quickly without getting lost in syntax or framework details unfamiliar to you.

   For example: “You're given a Java implementation of a merge-sort variant that produces wrong results on certain input patterns. Your debug approach would be to: describe one concrete test case that would fail and show how you'd trace execution.”

2. 2.Competitive programming problem analysis

   Evaluating AI solutions requires spotting algorithmic flaws, suboptimal trade-offs, and missed edge cases before they enter training data.

   For example: “An AI generated a valid O(n log n) solution to a sorting problem, but the code uses 3x the memory it needs. How would you annotate this for the training feedback, and why does it matter?”

3. 3.Writing actionable model feedback

   Models learn from your explanations; vague or wrong feedback teaches bad patterns; clear, specific feedback trains them correctly.

   For example: “An AI's code is correct but has a subtle off-by-one error in boundary handling that only surfaces on max-size inputs. Write the feedback you'd include in the training annotation.”

4. 4.Sustained productivity under output targets

   Payment depends on task completion; you need to manage your workflow to meet minimums without sacrificing quality.

   For example: “You're expected to submit 5-8 tasks per week. Two of your early submissions are rejected for unclear annotations. Describe how you'd adjust your process to meet both speed and quality.”

A task you may get

Given a competitive programming problem and an AI-generated solution with at least one subtle bug, identify the bug, write a test case that exposes it, and provide training feedback explaining the error and the fix.

How to prepare

• Solve 2-3 competitive programming problems (LeetCode medium or Codewars 4-kyu level) and time yourself to get comfortable with your speed
• Review your 2 best open-source contributions and prepare to discuss design choices and trade-offs you made
• Practice writing concise, specific technical feedback on code you read, focusing on clarity for someone unfamiliar with your style

The facts

Pay

$50–100/hr

Open to

Bangladesh, Hong Kong, India, Indonesia, Japan, Kazakhstan, Kyrgyzstan, Malaysia, Pakistan, Philippines, Singapore, Sri Lanka, Taiwan, Thailand, Uzbekistan, Vietnam, Austria, Belarus, Belgium, Denmark, France, Germany, Greece, Italy, Netherlands, Portugal, Russia, Spain, Switzerland, United Kingdom, Argentina, Brazil, Chile, Colombia, Mexico, Peru, Algeria, Bahrain, Egypt, Iraq, Jordan, Kuwait, Lebanon, Libya, Morocco, Oman, Palestine, Qatar, Saudi Arabia, Tunisia, United Arab Emirates, United States, Canada, Nigeria, Kenya, South Africa, Ghana, Ethiopia

Field

Software Engineering

Role type

Expert

Posted

9/14/2026

Places left

100

We wrote this page from the public micro1 listing. It may be out of date, so read the full posting before you apply.