Senior Software Code Reviewer
Vetto
WorldwideremotePosted 7 days ago
Skill Required
Code-ReviewSoftware-Quality-AssuranceSoftware-EngineerAI-Data-ValidationTechnical-AssessmentsCode-Review-AnalystSoftware-Quality-EngineerSenior-Software-Engineering-AnalystSenior-Software-Test-Engineersoftware engineeringEngineeringJavaScriptTypeScriptbuildingwrittenTestNGPythondesignDesign PatternsJavaRustGenerative AITDDandGoAIMachine LearningFulltime
Key highlights
- Compensation: $100 per hour (task‑equivalent reference), variable by complexity
- Target experience level: Mid/senior‑level software engineers
- Must have strong command of at least two languages: Python, JavaScript/TypeScript, Go, Rust, or Java
- Task‑based work: each review takes about 30 minutes
- Selection is fully asynchronous with no traditional interviews
- Role serves as a technical quality gate for AI model training
Role overview
Vetto, a tech company building high‑quality datasets for AI, is looking for experienced software engineers to perform technical review and validation of coding tasks used to train large language models. The role focuses on assessing whether generated code correctly solves the given prompts, evaluating the robustness and correctness of associated tests, and applying strict technical standards to ensure data quality for AI systems.
Responsibilities
- Review and analyze generated code against the original software engineering prompt
- Evaluate whether the coding task itself is clearly and correctly defined
- Validate whether tests accurately reflect whether the problem has been solved
- Identify gaps, ambiguities, false positives, and false negatives in test suites
- Determine whether a solution that passes the tests genuinely solves the underlying problem
- Apply strict technical criteria and quality standards consistently across tasks
Requirements
- Mid/senior-level software engineer with real professional experience
- Proven professional experience in software development (production environments)
- Strong command of at least two of the listed languages (Python, JavaScript/TypeScript, Go, Rust, Java)
- Experience reviewing and evaluating code written by other engineers
- Solid understanding of automated testing — how tests validate (or fail to validate) behavior
- Experience contributing to or working with open source projects
- High attention to detail and strong technical judgment
- Comfortable working fully in English (reading and writing)
Nice to have
- Experience with test‑driven development (TDD) or test design
- Familiarity with large or complex codebases
- Background in AI, ML, or data‑centric projects
Additional details
- About Vetto: Vetto is a tech company focused on building and scaling high‑quality datasets for artificial intelligence systems, working at the intersection of human expertise and AI to ensure models are trained on technically accurate, well‑defined, and realistic data.
- Our projects support the training and evaluation of Large Language Models (LLMs), where technical rigor and correctness are non‑negotiable.
- About the Project: This project focuses on the technical review and validation of coding tasks used to train AI models. Automated code is generated in response to software engineering prompts, and your role is to evaluate whether that code is correct and truly solves what was asked.
- Core questions for each task: Is the coding task technically well‑defined? Does the generated code actually solve the problem? Are the associated tests robust, correct, and aligned with real‑world software behavior? Tests are treated as the mechanism of truth; mistakes propagate at scale into AI systems, making error criticality high.
- Languages: Tasks involve code written in Python, JavaScript/TypeScript, Go, Rust, and Java. Strong command of at least two of these languages is required; breadth across languages is a plus.
- Project Details: This is expert, task‑based technical work focused on analysis, validation, and judgment — not code production. Each task takes approximately 30 minutes and is reviewed under continuous QA and calibration.
- Compensation: $100 per hour (task‑equivalent reference), varying based on task complexity and approved volume.
- Selection Process: Fully asynchronous; candidates are evaluated via background, screening responses, and a short technical exercise focused on code review and test validation. No traditional interviews are required.
- Final Note: The role is about technical judgment, rigor, and responsibility, not writing more code. It is suited for those comfortable challenging problem definitions, questioning tests that "pass but are wrong," and acting as a technical quality gate.
- Originally posted on Himalayas