Remote AI training and data labeling jobs

Experienced engineers review what AI coding agents produce in real repositories, build rubrics and label preference data. $100-150/hour, 20-40 hours a week, about 3 months, open to North America only, with at least 6 hours of Pacific Time overlap.

Tabletop and strategy-game experts build and grade rule-based reasoning tasks for AI models. Contractor assignment of 2 months, 20+ hours a week with PST overlap, take-home assessment. No rate published.

Engineers with 3+ years build coding datasets, correct and refine AI-generated code across Python, JavaScript, C/C++, Java, Rust and Go, and design automated verifiers. One month contractor term with possible extension, 10 to 40 hours a week, for US, Canada and Western Europe. No rate published.

Go engineers with 3+ years review and validate AI-generated code, resolve hard coding issues and help improve a dialog-agent system, with some work on large public GitHub repos. Contract; duration and hours flexible; two interviews. No rate published.

Senior engineers from top-tier product companies curate code, write solutions and corrections, and evaluate AI-generated code across Python, JavaScript, C/C++, Java, Rust and Go. Contract, 10 to 40 hours a week, 1 month with possible extension. No rate published.

Tech-lead-level engineers triage GitHub issues, containerize and set up popular open-source repos, assess test quality and run codebases to see how LLMs fix bugs. Contract, 20 hours a week, 3 months, some PST overlap. No rate published.

Turing contract for C and C++ developers to review and validate AI-generated code, improve public GitHub repositories and help train LLMs. Leadership experience wanted; contract length and hours flexible.

Turing contract for Java engineers who write training code, run evals, rank model answers and contribute to SFT and RLHF. 1 month, 20 to 40 hrs/week, 4 hrs overlap with PST.

Turing contract for JS/TS developers who write code to train AI models, run evals, rank model responses and contribute to SFT and RLHF. Node, React, Nest, Angular or Vue. 1 month, 20 to 40 hrs/week.

Python and JavaScript/TypeScript developers evaluate and rank model responses, write code and rationales for supervised fine-tuning and RLHF, and run evals for AI labs. One-month contractor assignment at 20, 30 or 40 hours a week with PST overlap. Docker required. No rate published.

Python developers with 3+ years write and review code for LLM training data: SFT responses, RLHF comparisons and model evaluations for foundational AI labs. Remote contract, 20, 30 or 40 hours a week, one month initial term. No rate published.