Strange Loop

InferenceQL: AI for data engineers, without the math

Can we empower data engineers to use AI to explore, monitor, clean, and predict data streams, without having to learn math?

InferenceQL is a new open-source AI platform for semi-structured data that is being developed at MIT. InferenceQL users can build models using automatic model discovery, then query these models using a simple, SQL-like API. InferenceQL also provides a spreadsheet interface with built-in data visualization. InferenceQL runs on both JavaScript and the JVM, so it can drive interactive data experiences in web pages and also be part of enterprise data pipelines. Models inside InferenceQL are represented as probabilistic programs in Metaprob, a new probabilistic programming language embedded in Clojure/Script. Experts can thus read and customize models using probabilistic programming techniques. The InferenceQL team is working on an open-source prototype suitable for industry and civic use in data engineering, analytics consulting, and data journalism, as well as scientific data analysis and research in probabilistic programming, causal modeling, and probabilistic expert systems.

InferenceQL was made by some of the same people behind BayesDB and Empirical Systems, first presented at Strange Loop in 2015. Unlike BayesDB, InferenceQL is not a database, but instead can be embedded alongside data tables, data streams, and client-side web applications. We are just starting to seek external beta testers, contributors, and volunteers.

Ulrich Schaechtle

Ulrich Schaechtle


Ulrich Schaechtle is a research scientist at MIT. He leads the research engineering efforts around InferenceQL. Ulrich holds a PhD in computer science from Royal Holloway, University of London, as well as an MSc in computing from Imperial College London and a BSc in applied congitive sciences from the University of Duisburg-Essen. Ulrich leads an MIT team for the DARPA Synergistic Discovery and Design (SD2) programs. He has publications in major conferences and journals for programming languages and AI. He currently works on applications of InferenceQL to data journalism, psychiatry, and synthetic biology. In 2018, Ulrich was selected by DARPA as a DARPA Riser, one of 50 early career scientists developing breakthrough technologies.