Custom and off-the-shelf data collection services including remote, on-site, device-based, and location/POI collections; supports image, video, speech/audio, text, documents, and location data, with preparation/annotation available via ADAP.

Custom data collection via remote, on-site, device-based, location/POI, and off-the-shelf dataset options, with workflow design and delivery through Appen’s platform and mobile app.

Enterprise AI data platform that merges automation and human oversight to manage data preparation and model evaluation workflows (annotation, classification, preference scoring, A/B testing, user testing, red teaming, benchmarking) across multiple modalities.

A flexible enterprise AI data platform that merges automation and human oversight to manage data preparation and model evaluation workflows across modalities (text, image, audio, video, 3D/4D), including annotation, classification, preference scoring, and evaluation methods like A/B testing, benchmarking, and red teaming.

Human-expert powered sourcing, curation, annotation, and evaluation of high-quality training datasets across modalities (text, image, audio, video) including hard-to-find and niche data requirements.

Human-expert powered data sourcing, curation, annotation, and evaluation services to produce high-fidelity training datasets for deep learning and traditional AI applications across modalities and industries.

Managed annotation services for text, audio, image, video, and multimodal datasets (e.g., sentiment, intent, NER, transcription, object detection, tracking, event detection) supported by Appen’s crowd and tooling.

Services for LLM data creation and improvement including supervised fine tuning, human preference ranking (RLHF/DPO), evaluation & A/B testing, red teaming/model safety, and RAG data preparation.

Services to support LLM development and enterprise customization, including supervised fine-tuning datasets, RLHF/DPO preference workflows, RAG data preparation, red teaming, and LLM evaluation and A/B testing.

Licensable catalog of ready-to-use datasets across audio, image, video, text, and location data, described as spanning hundreds of datasets across many languages and countries.