Kili Technology
Kili Technology provides a complete platform for building trustworthy, high-quality datasets used to train, fine-tune and evaluate AI/ML models. Its data-centric AI tooling turns raw image, video, text, document, audio and geospatial data into structured, reviewable training data through collaborative workflows.
At a Glance
- AI-powered scale-ups and data-science teams
- Large enterprises and multi-project AI operations
- Banking and insurance
- Manufacturing and electronics inspection
- +4 more
AI Tools by Kili Technology
(1)
Kili Technology
Data Labeling Platform for ML
Discussions
No discussions yet
Be the first to start a discussion about Kili Technology
Latest News
January Product Update: Precision Meets Productivity in AI Data Labeling — multi-select/copy/paste/drag annotation operations, private-beta node editing, enhanced GeoJSON exports and SDK configuration support.
What’s New on Kili — Key Enhancements for Geospatial Projects, including richer geospatial annotation and external map-layer capabilities.
The Evaluation Gap: Why AI Breaks in Reality Even When It Works in the Lab — Kili’s discussion of production AI evaluation and data-quality gaps.
What Works in Geospatial AI: An Expert Analysis — Kili’s analysis of effective geospatial AI workflows and dataset practices.
Products & Services
Enterprise platform for annotation, data curation, quality review, model-assisted labeling, dataset iteration and human evaluation across image, video, text, PDF, audio and geospatial data.
Configurable ontologies, classification, object/entity detection and relations, transcription, named-entity recognition, AI-assisted pre-annotation, consensus, honeypots and review workflows.
Multi-project workflow orchestration for distributed annotation teams, with role-based access, automatic task distribution, asset locking, quality gates and per-annotator analytics.
SOC 2 Type II, ISO 27001:2022 and HIPAA-certified workflows with project-level isolation, audit trails, granular roles, SaaS, Azure Marketplace and on-premise/Kubernetes deployment options.
Market Position
Kili positions itself as an enterprise-grade, data-centric AI platform spanning annotation, quality management, evaluation and workforce operations rather than a single-modality labeling tool. Its differentiation is breadth across modalities, collaborative multi-stage QA, APIs/SDKs, enterprise compliance and cloud-to-air-gapped deployment. Comparable competitors include Labelbox, Scale AI, Encord, SuperAnnotate, Dataloop, CVAT and Label Studio.
Leadership
Founders
Edouard d’Archimbaud
Co-founder and CTO. Previously worked at BNP Paribas, where he built one of Europe’s advanced AI labs from scratch; he led the technical vision around faster, higher-quality annotation and data preparation.
François-Xavier Leduc
Co-founder and former CEO/publication director. He focused on turning the insight that AI teams should prioritize data quality into a company and commercial platform.
Executive Team
Mathieu Azar
Chief Executive Officer
Became CEO in 2025; his prior experience includes Lucca, where he spent several years and was involved in the company’s growth and acquisition journey.
Edouard d’Archimbaud
Co-founder and Chief Technology Officer
Previously built an advanced AI lab from scratch at BNP Paribas; co-founded Kili to develop data-centric AI infrastructure and annotation technology.
Founding Story
Kili began as an idea in 2018 while Edouard d’Archimbaud was building an AI lab at BNP Paribas and François-Xavier Leduc was focused on turning powerful insights into a company. They started Kili because they believed data quality, rather than continued model preparation alone, was the major barrier to better AI. The platform went live in July 2020, and early customers renewed by the end of 2020.
Business Model
Revenue Model
B2B SaaS subscriptions for the annotation/evaluation platform, with usage and capacity-based limits, plus custom enterprise contracts, professional services, dedicated labeling services and deployment/security add-ons.
Pricing Tiers
One team seat; 100 assets for text, documents and images; 5 assets for video and satellite imagery; AI-assisted labeling; no credit card required.
Up to 20 team seats and 50,000 assets, API and Python SDK, adapted support and accessible professional services.
Grow features plus custom seat allocation, custom terms, SSO and advanced security features; on-premise deployment and other add-ons are available.
Target Markets
- AI-powered scale-ups and data-science teams
- Large enterprises and multi-project AI operations
- Banking and insurance
- Manufacturing and electronics inspection
- Defense and government
- Healthcare and life sciences
- Large-scale AI training and fine-tuning
- Model evaluation and human-in-the-loop review
- Multi-project annotation programs
- Computer vision and electronics quality inspection
- OCR, intelligent document processing and KYC/document extraction
- NLP, LLM and named-entity workflows
- SAP
- Safran
- Louis Vuitton
- Capgemini