Fiddler AI

Fiddler AI is an AI observability and governance software vendor founded in 2018 by Krishna Gade and Amit Paka, headquartered in Palo Alto, California. It targets a problem that model risk functions at banks know well: once credit, fraud and pricing models — and now LLM-based agents — run in production, their behaviour drifts, every decision must be explainable to risk committees and supervisors, and failures are hard to trace back to a cause.

The product is positioned as a control plane for AI and spans ML Observability for predictive models, Agentic Observability and Continuous Evaluations for LLM applications, Policy Enforcement guardrails, and an AI Governance, Risk Management and Compliance module, with more than 100 pre-defined metrics plus custom metrics and proprietary Fiddler Centor evaluation models. Named customers include Mastercard, Ally, LendingPoint, American Family Insurance, CSAA, Nielsen and the US Navy. In January 2026 Fiddler raised a USD 30 million Series C led by RPS Ventures, bringing total funding to USD 100 million.

1. ML Observability: continuous monitoring of drift, data integrity and performance for production credit, fraud and pricing models.

2. Explainability: attribution and root-cause analysis so individual model decisions can be justified to risk, audit and supervisors.

3. Agentic Observability: tracing and evaluation of LLM agents across their lifecycle, from development into production.

4. Continuous Evaluations: 100+ pre-defined metrics plus custom metrics, powered by proprietary Fiddler Centor evaluation models.

5. Policy Enforcement: runtime guardrails that block unsafe or non-compliant model and agent behaviour before it reaches customers.

6. AI Governance, Risk Management and Compliance module for regulated deployments in lending, cards, wealth management, fraud and trading.