Skip to content
One6
One6

Capability statement

Human data for AI training and evaluation

Human data for AI teams: training data collection, annotation, multilingual and speech data, and expert model evaluation from a global contributor network.

7,000+hours of contributor work delivered on client AI data projects

Core services

  • AI Training Data Collection: Voice, text, image, video and custom datasets collected from real people to your specification.
  • Data Annotation & Labeling: Image, video, text and audio labeling with written guidelines and layered quality review.
  • Multilingual AI Data: Native-speaker collection, transcription, localization and evaluation across languages.
  • Speech & Voice Data: Conversational and scripted speech, accents, emotion and professional voice talent.
  • Human Evaluation for AI Models: Domain experts and trained generalists rate, test and review model output.

Solutions

  • LLM Fine-Tuning Data
  • LLM & Model Evaluation
  • Speech & Voice AI
  • Multilingual Expansion
  • Computer Vision Data

Industries

AI & Technology · Healthcare & Life Sciences · Financial Services · Legal · Public Sector · Research & Academia

Delivery lifecycle

  1. 01 Specify. Requirements become a written specification and a guidelines document agreed with your team.
  2. 02 Qualify and train. Contributors pass a qualification test, then train on the project's guidelines.
  3. 03 Pilot. A small first batch you review before committing to volume.
  4. 04 Deliver and track. Regular batches with ongoing QA, tracked contributor performance and regular status updates.

Contributor network

  • Domain experts, native speakers, trained generalists and voice talent
  • Recruited to match each project's languages, domain and demographics
  • Skills test or trial task before joining a project
  • Training on each project's guidelines before production work
  • Accuracy and output tracked; contributors below the bar retrained or removed

Quality controls

  • Written guidelines agreed with the client and kept up to date
  • Reviewer calibration on shared examples
  • Gold-standard items, reviewer sampling and agreement monitoring
  • Quality results shared during the pilot

Communication

  • Regular status updates by email or scheduled calls
  • Project oversight scaled to the size and complexity of each program

Data rights and confidentiality

  • Data ownership and usage rights defined in the project agreement
  • Projects can run under NDA; contributors keep project materials confidential
  • Consent and usage terms agreed for recorded data before collection
  • Work in the client's platform or tooling proposed during scoping

Contact

hello@one6.ai

www.one6.ai/contact