AI Startups
The Institute for Experiential AI (EAI) at Northeastern University harnesses the power of artificial intelligence to address real-world challenges across various industries. We offer cutting-edge research, practical AI applications, and tailored AI solutions to drive innovation, improve efficiency, and foster responsible AI practices. Our interdisciplinary team collaborates with industry leaders to integrate AI into their workflows, ensuring impactful and ethical outcomes.
Cutting-edge legal research powered by generative AI. Sort through millions of cases, statutes, and regulations in minutes. All for less than a cup of coffee.
Jaxon is addressing the hallucination problem with a formal reasoning system that mathematically proves the output from LLMs is accurate. Large language models like GPT-4 "hallucinate", i.e., can fabricate responses. Based on R&D with the US Department of Defense, Jaxon's proprietary DSAIL (Domain-Specific AI Logic) technology turns natural language into structured code that gets run through a gauntlet of checks and balances. This ensures the AI's response meets all constraints and assertions before being returned - formal verification and validation of its output.
First Touch is an AI powered research and outreach platform that turns sales reps into super reps. First Touch is designed to help sales reps create highly personalized and effective outreach. The platform is an action layer that pulls in intent signals like podcasts, social posts, tech stack, recent news, hiring, funding and more, enabling sales reps to generate personalized outreach that resonates deeply with their target audience across the entire customer journey. First Touch is a google extension that lives within your existing SEP (no integration needed), Linkedin, and URL. First Touch aims to revolutionize outbound sales by focusing on delivering value at every interaction, fostering genuine human connections, and empowering sales reps to efficiently manage their outreach and engagement processes.
Oblio is building a comprehensive Data Lakehouse for Higher Education. Our platform integrates data from every campus department into a unified, AI-powered system. We don't just store data; we transform it into strategic action. Our AI analyzes patterns across the student lifecycle, from prospective applicant to alumni, generating personalized communication strategies. For admissions, this means tailored outreach that significantly boosts yield rates. For advancement offices, we unlock "Behavioral Chain Philanthropy," leveraging historical data to cultivate stronger, more lucrative alumni relationships. Our system is designed for ease of use, allowing any staff member to harness the power of advanced data analytics without specialized training. With Oblio, institutions gain a competitive edge in enrollment, student success, and fundraising - all through one intuitive platform.
AI platform for creating, deploying and managing scientifically rigorous, physics informed AI agents for autonomous control of industrial systems and assets. Our initial focus is on the energy sector where we are delivering breakthroughs increased profits while reducing green house gas emissions.
Human-like agents that handle inbound and outbound customer calls 24/7 and come pre-skilled to intake new clients, book appointments, follow up on leads, and much more.
Sway AI help companies of all sizes adopt AI predictive analytics to solve real business problems without big technical teams. We achieve this with customizable solutions and a no-code AI environment, enabling our customers to improve efficiency and accelerate growth.
ekai automates the messy middle of data operations to enable enterprises to enter the genAI age. Today, transforming fragmented data from applications and microservices contained across multiple systems into usable data structures costs millions and takes thousands of person-hours in data modeling, integration, and contextualization. ekai revolutionizes this process by automating complex data modeling tasks using proprietary AI technology. Our specialized algorithms are designed to handle the intricacies of data harmonization and contextualization, minimizing the need for extensive manual input, and streamlining the entire data transformation process - thus decreasing the time to build any genAI use-case for enterprises from months to minutes.
TrueFoundry is a self-hosted, cloud-native, platform as a service that automates all of the engineering and infrastructure orchestration needed to build and take Generative applications from PoC to Production. We are working with multiple enterprises like NVIDIA, Siemens, Merck, and CVS among 20 other Enterprises and Mid Market Tech Companies. In their Generative AI Journey, we have ensured that all applications built by these companies are: 1) Modular 2) Cost Efficient 3) Properly governed for Cost and Risks 4) Performant and scalable.
​
The founding team consists of Nikunj Bajaj
(www.linkedin.com/in/nikunj-bajaj-10476824/), Abhishek Choudhary (www.linkedin.com/in/abhishekch123/) and Anuraag Gutgutia (www.linkedin.com/in/anuraaggutgutia/). Abhishek and Nikunj were leading ML teams at Meta and realized that the horizontal platform layer present in the company helped AI and ML teams move with speeds that were not possible with the stack of any company that has not heavily invested in making such a platform. They wanted to democratizate the access to such a platform to all companies and hence built TrueFoundry.
​
The platform comes with multiple modules, such as LLM Playground (Experimentation layer: compare models, test prompts, etc.), LLM Gateway (Audit Logging, Governance, Cost Management, third-party API management), LLM Deploy (optimally Deploying any Open Source Model for Inference), LLM Fine-tuning, and LLM Pre-training, which ensure that companies can move much faster by using these core components and can focus only on building use cases to deliver value. We have also made production-ready, editable application bundles like RAG which companies use to build on top of and save time massively.​
​
Some of our customer case studies are available here: www.truefoundry.com/case-studies
Here is some of the love that our customers have shown us: www.truefoundry.webflow.io/wall-of-love
Demo: www.youtube.com/watch?v=oU1lxXFctmw
Website: www.truefoundry.com
Cost Savings through TrueFoundry:
Descrybe.ai is a new way to search for and understand case law. Our unique process leverages generative AI to make complex legal information more accessible to professionals and laypeople alike. We are laser-focused on easy access to case law research, by lowering cost (it's free!), increasing ease of use, and employing natural language search and summarization capability.
​
But working to increase access to the law is not the only thing that separates us from other legal search engines. We are also taking a different approach to generative AI. While our process does use OpenAI, we are not just another front-end for ChatGPT. In fact, we don't use ChatGPT at all. Our searchable summaries are stored in our own databases; cases are not summarized on-the-fly, nor is the internet searched to complete searches. This makes our tool incredibly fast, accurate, and far less likely to suffer from hallucinations. We also don't track users or have any way to associate searches with individuals, which helps our tool be less of a concern for issues around confidentiality.
​
While our approach could make any type of complex information more accessible, our first consumer-facing application is a legal search engine that anyone can access, use, and understand. The database currently has more than 3.3 million state and federal judicial opinions with more being added daily.
​
We are also working on several new features, some are live and some are in development. We recently rolled out a new feature we call "descrybe abstracts," which give users an easy way to get the gist of an opinion at a glance. We currently have more than 1.15M abstracts available (look for the sparkle icon next to the search result). We will also be rolling out our first paid product in early fall 2024. The product is tentatively called the "Cytationator" and will be a low cost and accessible way to check the accuracy of the citations included in a legal brief. We are also developing several other new features that will launch later in the fall.
​
Descrybe.ai has been summarizing opinions since February 2023, launched in July 2023, became a member of the Justice Technology Association in September 2023, became a sponsor of the Massachusetts Academy of Trial Attorneys in October 2023, was named a finalist in the Legalweek Leaders in Tech Law Awards (Enabling Access to Justice category) along with industry leaders Casetext (part of Thomson Reuters), Clearbrief, Relativity, and Reveal in November 2023 (Relativity won), was named a semifinalist in Bob Ambrogi's Startup Alley at ABA TECHSHOW 2024 in December 2023, became a member of the Institute for the Advancement of the American Legal System's Above the Line Network in February 2024, and was named a nominee for a Webby Award in the AI, Metaverse & Virtual - Public Service & Activism category in April 2024. In addition, our co-Founder Kara Peterson was recently selected for the prestigious American Bar Association's Distinguished Women of Legal Tech Award (one of only eight women in the United States selected for this honor).
​
We have enjoyed a warm welcome from leaders in the legal tech community, with many telling us that what we have built is powerful, unique, and potentially game-changing. We have users from across the globe, including sitting judges, law students, practicing lawyers, paralegals, legal librarians, and journalists. We have appeared in many prominent legal tech podcasts and publications including MIT Media Lab IdeaFlow with Dazza Greenwood, Notes to My Legal Self with Olga Mack, Conversations with LSNTAP (Legal Services National Technology Assistance Project) with Shellie Reid, AI in Real Life: Not Your Usual Take with Elena Gurevich and Alisa Sedneva, Kassi &... with Kassi Burns, Women and AI with Lana Dubinskiy, Reut Lazo, and Jenny Kay Pollock, Tech Thursday with Cheryl Wilson Griffin, and Pinecone's RAGBrag. We have several more appearances scheduled for the second half of 2024 including a second appearance on Notes to My Legal Self with Olga Mack and LawDriod Manifesto with Thomas G. Martin. We have also been covered by the media including a feature by leading legal tech journalist Bob Ambrogi in LawSites as well as several stories in Legaltech News by ALM and BostInno, among others. You can see all the coverage we have so far (and awards we have been nominated for) on our news page.
​
We are also participating in various AI events as invited guests or speakers. Recent examples include participating on Bob Ambrogi's AI in Practice panel at Suffolk University Law School's LITCon on April 8 and presenting at the selective Imagination in Action: Forging the Future of Business with AI Summit at the MIT Media Lab on April 18.
​
Descrybe has also named four renowned computational law/legal research experts as strategic advisors, namely Dazza Greenwood of CIVICS.com and the MIT Media Lab, Dr. Megan Ma of Stanford Law School and CodeX, Rebecca Fordon of The Ohio State University Moritz College of Law, and Sean A. Harrington from the University of Oklahoma College of Law.
​
Descrybe.ai was built in Greater Boston by Richard DiBona and Kara Peterson, a husband and wife team who share an interest in technology, the law, and the democratization of information. Rich brings deep experience in using emerging technology to solve complex problems, and Kara is a communications and marketing expert with a special focus on mission-driven organizations and social justice.