Use your STEM expertise to help improve large language models across scientific and technical fields. Project work may involve creating and solving challenging problems, writing domain-specific prompts, and judging model answers for accuracy and reasoning quality. The year-round fellowship offers remote, asynchronous project work for about 5 to 20 hours per week; assignments depend on project availability.
Lead Japanese quality reviews across AI training projects, checking model outputs and contributor work against project rubrics. You’ll give specific feedback, answer language questions, coordinate updates through Discord, and maintain resources such as Japanese style guides and QA trackers. This is a remote contract role with no immediate project; qualified candidates may be contacted when relevant work becomes available.
Assess how well a large language model handles Swahili tasks such as translation, summarization, dialogue, and grammar correction. You will also refine and annotate Swahili text, flag language gaps, and share detailed feedback with AI researchers. The project focuses on building high-quality language data and improving the model’s Swahili understanding and generation.
Legal Counsels contribute legal expertise to a customer project that helps train AI systems. The work draws on detailed legal analysis and strong written and verbal communication, including turning complex legal material into accessible guidance. No previous AI experience is required; legal expertise is the focus.
Create reference advertising edits and assess AI-generated video for training and evaluation projects. You’ll bring studio or agency editing experience to hands-on post-production, build pacing and continuity standards, and give specific feedback on model output. The work covers commercial and social content in several aspect ratios, with a 15-minute AI interview for shortlisted candidates and a 45-minute technical interview for those who advance.
Record a short script in your natural Romanian-influenced English accent, using a smartphone or computer with a microphone. Follow the supplied recording guide and submit your audio through the designated platform. The recording takes about 15 minutes or less, and the project pays $50 per submission. The project accepts 3 to 5 speakers for this accent, and people of all genders may apply.
As an independent contractor on an audio and language model evaluation project, you’ll assess Dutch samples for accuracy, fluency, naturalness, and overall quality. Your reviews may require close attention to grammar, pronunciation, tone, phrasing, and word choice, including whether spoken Dutch fits its context. Each task takes about 30 seconds to 4 minutes.
Review and evaluate data science documents used to train a language model for a foundational AI lab. You’ll assess whether each document’s content and structure match the materials data scientists encounter in their day-to-day work. The role is fully remote and pays $100 per hour; task volume varies, hours are not guaranteed, and most experts work under 10 hours per week. There is no weekly cap on hours or earnings. Start date: ASAP.
Contribute to short-term AI research projects with leading AI labs by applying your academic knowledge to improve large language models in specific fields and topics. Work remotely with no minimum hour commitment. Handshake provides brief training, and you must pass one assessment before starting. Successful project work may qualify you for future projects.
Guide quality across Korean AI training projects by reviewing model outputs and trainer or QA work against project rubrics. You will give specific feedback, resolve questions about Korean usage, and keep contributor guidance current through Discord and Google Docs. The role is remote and contract-based, with no immediate project; qualified candidates may hear about future assignments through the expert network.
Assess and improve a large language model’s performance in Singaporean Chinese (Simplified). You’ll review generated text across translation, summarization, dialogue, and grammar correction, refine language to fit local cultural and linguistic norms, and prepare annotated data. Your findings will help AI researchers address model errors and strengthen Chinese text generation and understanding.
Work on an audio annotation project that helps train AI systems to recognize Belgian and Swiss French. You’ll assess conversational recordings, label audio and speaker details, and correct speech recognition transcripts using regional conventions. The role suits someone with strong knowledge of local French varieties and experience in language-related work such as teaching, transcription, editing, translation, or journalism.
Guide architecture and product decisions for an AI-powered youth sports media platform that turns athletic video into stories and recognizes athletes. The company has a polished prototype, a roadmap, and network design options, and is validating interest with pilot customers before development begins. Start with a paid two-week trial of 10–15 hours, then move to an anticipated six-month engagement at about five hours per week, with room for more hours and a possible longer-term CTO role.
Record a short script in your natural Tamil-influenced English accent, using a smartphone or computer with a microphone and the provided step-by-step guide. The recording takes about 15 minutes or less, with a flat payment of $50 per submission. The project seeks 3 to 5 speakers for this accent, and contributors of all genders may apply. Imitated accents do not qualify for compensation.
Evaluate AI-generated technical and customer-facing responses, drawing on language fluency, coding knowledge, and customer service judgment. You’ll check response quality and JSON structure against project guidelines. The available language cohorts are Dutch (Netherlands), Welsh, French (Canada), German (Germany), Italian (Italy), Portuguese (Brazil), Greek, English (US), and English (UK).
Help train a foundational AI lab’s language model by creating, assessing, and refining professional product presentations and slide decks. Work remotely on your own schedule, with an ASAP start. Pay is $80 per hour, but hours depend on client task volume and are not guaranteed; most experts work under 10 hours a week, with no weekly cap on hours or earnings.
The year-round Handshake AI Fellowship brings engineering specialists into project-based work that helps improve large language models in expert fields. You’ll create subject-specific prompts, assess model responses, and research topics with AI. Work remotely on your own schedule for about 10 to 20 hours each week; project placement depends on availability.
Review AI-generated Swedish and create reference answers that show accurate grammar, natural phrasing, and appropriate tone. You will compare model responses, check facts and cultural context, and explain language issues in written feedback. The work calls for native or near-native Swedish, C1-level English, and at least three years of language-focused professional experience.
Location and eligibility
Remote · Finland, Sweden
Engagement type
Contract
Compensation
Up to $55/hr
swedish
english
copyediting
Language & LinguisticsView job
Showing 20 active roles of 419. Dates reflect what the source published, not when AIWorkJobs last synced.
Sources: ten specialist AI work platforms. AIWorkJobs does not verify every claim made by a source. Report a listing that looks wrong through the contact page.