Researcher

Jason Mars

Jason Mars is a computer scientist whose AI work spans conversational assistants and programming systems for AI applications. He began at the University of Virginia in 2006, earned a computer science PhD there in 2012, worked as a Google research intern from 2008 to 2011, and received Google's 2010 Compiler Technology Fellowship. After teaching at UC San Diego, he joined Michigan's faculty, co-directed Clarity Lab, where Sirius was developed, and co-founded Clinc with Lingjia Tang. His 2022 paper presented Jaseci and Jac, and later papers introduced Meaning-Typed Programming and Semantic Context Annotations for language-level LLM integration. As of October 2026, he is a computer science professor at Michigan, directs Clarity Lab, and is President of Jaseci Labs and co-founder of Myca.ai.

Now at
University of Michigan, Jaseci Labs and Myca.ai
Born
1983

Years link to their sources.

Timeline

  1. 2006–2012

    A PhD at Virginia

    He begins at the University of Virginia in 2006 and earns his computer science PhD there in 2012.

  2. 2008–2011

    Google research internships

    From 2008 to 2011, he works as a graduate research intern at Google in Mountain View, contributing to the design and prototyping of algorithms and projects.

  3. 2010

    A compiler technology fellowship

    In 2010, he receives the Google U.S./Canada Fellowship in Compiler Technology.

  4. By 2015

    Clarity Lab and Sirius

    Before joining Michigan, he is a professor at UC San Diego. By 2015, he is an assistant professor of computer science and engineering at Michigan and co-director of Clarity Lab, where Sirius is developed.

  5. 2015–2017

    Clinc and its assistants

    Mars and Lingjia Tang are Clinc cofounders. Their student team debuts Lucida, an open-source AI assistant, in 2015. In 2017, Clinc raises $6.3 million in a Series A round and launches Finie, its voice-controlled banking platform.

  6. 2022

    The Jaseci stack

    His 2022 paper presents Jaseci as a production-grade system stack and Jac as its corresponding programming language.

  7. 2024–2025

    Meaning-Typed Programming

    In 2024, he coauthors Meaning-Typed Programming (MTP), which abstracts LLM integration through language-level constructs. A 2025 paper he coauthors presents Semantic Context Annotations (SemTexts), which let developers embed natural-language context into Jac program constructs for prompt generation.

  8. As of October 2026

    Jaseci research and leadership

    Since August 2025, he serves as principal investigator for NSF's POSE Phase 1 project to build an open-source Jaseci ecosystem through July 2027. As of October 2026, he is a computer science professor at Michigan, directs Clarity Lab, and is President of Jaseci Labs and co-founder of Myca.ai.

Career

  1. NowUniversity of MichiganAssistant professor of computer science and engineering
  2. NowJaseci LabsPresident
  3. NowMyca.aiCo-founder
  4. Until 2015University of California, San DiegoProfessor
  5. From 2015ClincCo-founder
  6. 2008–2011GoogleResearch intern

Education

  1. 2012University of VirginiaPhD

Sources and credits

Photo: We have no published photo of Jason Mars yet. Send us a free-licensed photo.

  1. Jason Mars - Wikipedia en.wikipedia.org. Checked 21 October 2026.
  2. Google Student Blog: After the award: students and mentors students.googleblog.com. Checked 21 October 2026.
  3. _bio_ - Jason Mars jasonmars.org. Checked 21 October 2026.
  4. Clarity Lab at University of Michigan launches platform that allows you to build your own Siri - Jason Mars jasonmars.org. Checked 21 October 2026.
  5. Look out, Siri: This Ann Arbor startup is revolutionizing artificial intelligence – Concentrate concentratemedia.com. Checked 21 October 2026.
  6. Clinc closes $6.3 million funding round, eyes major clients for AI financial assistant – Concentrate concentratemedia.com. Checked 21 October 2026.
  7. The Case for a Wholistic Serverless Programming Paradigm and Full Stack Automation for AI and Beyond -- The Philosophy of Jaseci and Jac arxiv.org. Checked 21 October 2026.
  8. MTP: A Meaning-Typed Language Abstraction for AI-Integrated Programming arxiv.org. Checked 21 October 2026.
  9. Prompt Less, Smile More: MTP with Semantic Engineering in Lieu of Prompt Engineering arxiv.org. Checked 21 October 2026.
  10. Jason Mars - National Science Foundation nsf.elsevierpure.com. Checked 21 October 2026.