The MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) is known for pioneering revolutionary approaches to computing and attracting top talent in tech.
With 10 Turing Award winners and past innovations like the fax machine, email, robotic arms, and text-to-speech technology, it’s no surprise that new innovations continue to emerge from MIT’s largest interdisciplinary research lab.
CSAIL students and researchers go on to make deep and far-reaching impacts in industry and beyond. Many of their groundbreaking discoveries are transformed into a product or platform that improves lives. Entrepreneurs interested in keeping their fingers on the pulse of innovation enjoy staying engaged with the MIT community through the Startup Connect program with CSAIL Alliances.
Below, we highlight several exciting startups that were founded by CSAIL graduates and researchers.
The future of pathology is paved with diagnostic tools that help pathologists to better diagnose and treat diseases. PathAI was co-founded in 2015 by MIT PhD graduate Aditya Khosla, whose work in CSAIL’s Computer Vision Group helped him build leading AI-powered technology for the pathology laboratory.
PathAI uses machine learning models to analyze digitized pathology images, bringing new insights into diseases and creating opportunities for more accurate and predictive diagnostics to patients, as well as making the process faster and cheaper. The startup works with board-certified pathologists to continue their research and development and improve patient outcomes.
Composable Analytics offers an advanced analytics ecosystem for the enterprise level. The Composable DataOps platform is built with intelligent methods that automate and augment user actions when building data-driven applications, helping business users intuitively architect enterprise-scale data intelligence solutions.
Through their research at MIT CSAIL, the initial prototype for a dataflow-based orchestration engine for big data was developed by Composable co-founders Andy Vidan and Lars Fiedler, and a technical team composed of researchers in CSAIL and MIT Lincoln Laboratory. Composable still fosters a “strong MIT culture of excellence, creativity, and innovation” within their team.
In today’s increasingly intelligent world, managing model versions and deploying them for production is often challenging. Data science teams looking to track, version, and audit models used across products can now take advantage of the Verta MLOps software ecosystem to manage models across their lifecycle.
Verta was founded by MIT graduate Manasi Vartak, who developed the open-source project ModelDB while working in the Database Group in CSAIL. ModelDB is a database that tracks versions of machine models. Verta builds upon this technology to track model versions and also operationalize them, helping data scientists put the models into production and, once deployed, ensure the models are still accurately reflecting the latest data.
For more information about Verta: https://www.verta.ai/.
SAIL uses Federated Learning to help companies and hospitals share sensitive data and valuable algorithms by protecting them from misuse during analysis. With the SAIL platform and its federated network, researchers can rest easy knowing that they can train predictive models and conduct research while preserving data privacy.
Co-founded by Professor Manolis Kellis of MIT CSAIL, the startup builds upon new encryption technology patented at MIT. This technology allows hospitals to fully control how, where, and when their data and algorithms are used, all with a simple user interface for researchers.
For more information about SAIL: https://secureailabs.com/.
Have you ever wanted to see what rearranging your office or living room would look like, without doing all the heavy lifting? Hosta uses AI to help you assess spaces, converting photos of spaces into intelligent 3D models.
While at MIT CSAIL, co-founder Rachelle Villalon focused on the intersection of AI and architecture in computer vision to develop this technology. Extracting rich information from images to generate materials, floor plans, production-ready CAD models, and more, Hosta saves time and money for anyone planning indoor spaces.
For more information about Hosta Labs: https://hostalabs.com/beta/.
Leela AI is a software layer that relies on resilient intelligence, or common-sense reasoning and decision-making for robots. Like humans, Leela can teach herself about her environment, exploring her environment without rewards or supervision. Leela can also communicate with a deeper level of understanding of meaning than current systems.
This software layer is based on research from the MIT AI Lab, on computation models of child learning combined with neural network image and sensor processing. Founders Henry Minsky, Dr. Cyrus Shaoul, and Dr. Milan Minsky met while students at MIT and have pursued AI research throughout their careers, continuing to push what is possible by developing technology that can better reason and learn. The company is currently working on a product prototype that will have robust machine learning applications.
For more information about Leela AI: http://leela.ai/.