Valohai

Valohai is an MLOps platform for building, evaluating, and operating AI products that combine large language models with specialized models trained on an organization’s data. Founded in 2016, the company helps teams manage the workflows surrounding AI, from experimentation and model comparison to deployment, retraining, and ongoing quality and cost monitoring.

The platform supports pipelines that connect scripts, APIs, and other workflow steps, with caching, parallel execution, conditional quality gates, and human approvals. Its dataset management tools provide versioning, file-level updates, and reusable caching, while automatic lineage links data, executions, and model versions for reproducibility and auditability. Valohai also offers tools for comparing LLM configurations, prompts, retrieval strategies, and models against real datasets, with quality, latency, and cost results. Teams can run workflows on their own cloud infrastructure, across multiple clouds, or in on-premises and hybrid environments. Valohai provides integrations and resources including documentation, training courses, an applied LLM certification, and support for agent-assisted project migration.