Home / Founders / Directory / Krishna Gade
Krishna Gade
Co-founder and CEO · Fiddler AI
Explainable AI Model monitoring
Gade wants every company running AI in production to be able to answer the question his Facebook team struggled with: why did the model do that?
Krishna Gade is co-founder and CEO of Fiddler AI, a company that monitors and explains machine learning models and AI agents once they are running in production. An engineer who came to the United States for graduate school in data mining, he worked on search, data infrastructure, and ranking systems at Microsoft, Twitter, Pinterest, and Facebook before starting Fiddler in 2018.
Data mining and search engines
Gade did his graduate work at the University of Minnesota, where he built clustering algorithms for large collections of text. He told James Le that he was lucky to study under pioneers of data mining and graph partitioning there.
He then worked on Microsoft’s Bing search engine and later led data engineering teams at Twitter and Pinterest as both companies scaled.
Explaining News Feed at Facebook
At Facebook he led a team that built debugging and transparency tools for News Feed, where hundreds of models combined to decide what each person saw. “It was very difficult to answer questions like ‘Why am I seeing this story?’ or ‘Why is this story going viral?’” he told Amazon Science.
Founding Fiddler
In late 2018 he left to start Fiddler in Mountain View, California, with Amit Paka, a former colleague and classmate who had worked on Samsung’s shopping apps, and Manoj Cheenath. “I realized this wasn’t a problem that just Facebook had to solve, but that it was a very general machine learning workflow problem,” he said.
The name comes from the way Fiddler lets a user change a model’s inputs, such as the size of a loan request, and see how the prediction moves. Gade describes it as working through a detective plot to find the root cause of a decision.
Accountability for algorithms
Gade has argued that companies cannot hide behind their models. Pointing to a credit card launch where couples with similar incomes received very different credit limits, he said, “We can’t abdicate our responsibility to an algorithm.”
