Latest AI Training Jobs and Career Resources

rustLabs pays vetted domain experts to build reinforcement-learning data and evaluations for AI labs (especially smaller labs), converting specialized knowledge into training data across software engineering, finance, ML, science, legal, and medical. Experts apply with proof of expertise, complete a calibration task, then pick projects and work remotely by the hour.

Pasiflora AI (launched April 2026, Eagan, Minnesota) connects AI labs with credentialed domain experts (PhD, MD, JD, CFA) to produce specialized training data that general crowdsourcing cannot. Every approved expert completes its 'Bloom Program' onboarding before being staffed, and work is matched by credential level. Led by CEO Kari Olson.

Pila8 is a mobile contributor app by Y Combinator-backed BeSimple AI (founded 2025, Redwood City) that pays users to record short voice and video clips used to train AI speech models on accents, dialects, and tonalities. Contributors read prompts, record and submit audio, and are paid per approved minute; the app also offers LLM red-teaming and evaluation tasks.

Thoth AI is a Singapore-headquartered AI data-services company providing multilingual training data across 200+ languages and 100+ countries, plus content moderation. It supports the full data lifecycle - collection, cleaning, and annotation of images, video, 3D point clouds, text, and audio - operating as a managed-services partner rather than a self-serve gig marketplace.

Luel (Y Combinator W26) is a rights-cleared marketplace and collection engine for multimodal AI training data. Enterprises submit a dataset spec and Luel recruits from a 500,000+ global contributor network across 96 countries, runs multi-stage QA, and delivers licensed, audit-ready datasets. Founded by William Namgyal and Inigo Lenderking; backed by Lightspeed Venture Partners.

Labela is a platform where medical experts get paid to label medical data and evaluate medical AI outputs on a flexible schedule, positioning itself as a healthcare/life-science domain-expert community for AI training. It targets credentialed medical professionals rather than generalist annotators.