I co-founded Graphistry in Q1 2014: We help analysts 100X investigations over operational data. Graphistry builds on our insights around end-to-end data analysis by combining visual workflow automation, visual GPU computing, and visual analytics. We currently work with F500, tech companies, and federal agencies, especially in cybersecurity investigations (IR, hunt, & intel) and anti-fraud (account takeover, bots, AML, etc.). Due to the power of our GPU graph technology and its usability, we also assist other graph/visibility projects, such as in network visibility, misinformation & bot hunting, cancer research, and anti-human-trafficking. Our initial project on GPU dataframes (an NSF SBIR) has been adapted by our Nvidia collaborators as RAPIDS and Apache Arrow, both of which are top data projects of 2019. We contribute to those and other supercomputing-democratization efforts, including helping start the GOAI: GPU Open Analytics Initiative.

Leading to our startup, I worked with Ras Bodik in the UC Berkeley Par Lab on the Superconductor GPU language for big data visualization. We also built the first parallel web browser, which led the way to projects like Servo by Mozilla. On the side, I explored the sociology of programming languages, built Proofist, a crowdsourcing proofreader, and researched language-based security techniques for the web similar to the subsequent CSP standard and JavaScript proxy objects.

Even earlier, I was a member of Brown PLT with Shriram Krishnamurthi. While there, we did the first functional reactive web language Flapjax and began my dalliance with language-based security. I have also worked at a startup, Microsoft Research, Macromedia, Adobe's Advanced Technology Labs, and WBRU, a commercial radio station.

My following languages paint a brighter future for the web:

Finally, causing some excitement, see our work on Socio-PLT: Sociology for Programming Languages!


(Out of date, sorry!)


(Out of date, sorry!)

Awards & Support

The parallel browser project was one of the motivating applications for the Parallel Funding Computing Laboratory (over $10 million from Intel, Microsoft, the State of California, and others) and we received individual grants for the parallel browser group (Samsung, Nokia, Intel, and Google.)


Ph.D., Berkeley 2013: My thesis examined the synthesis of parallel attribute grammar evaluators. I explored two applications: the first end-to-end multicore web browser, and a interactive data visualization scripting language with automatic GPU acceleration. It showed how to formalize, mechanize, and automatically parallelize the semantics of layout languages. The basic insight is that, by specifying layout widgets in a restricted constraint language, a synthesizer can schedule it as data parallel patterns over trees. In achieving it, we ran into interesting language constructs like partial parallel schedules, and new parallel algorithms like staged dynamic GPU memory allocation and speculative SIMD traversals.

At Berkeley and earlier at Brown (Sc.B. in 2007), I also researched language-based security, programming language adoption, and functional reactive web programming.

Parallel Browser
& GPU Data Viz

Languages, compilers, and algorithms


Social factors of language design


Dynamic languages


Analyzable and scriptable policies


    li>UPCRC (MSR Faculty Summit): The Parallel Browser: Architecture and Synthesis over Parallel Patterns (July, 2012)
  1. Microsoft Research (China): Parallel Browser / Socio-PLT (short talk, June, 2012)
  2. PLDI+ECOOP 2012 (SRC): Synthesizing and Parallelizing Layout Languages (finalist / short talk, June, 2012)
  3. Mozilla: FTL: Synthesizing a Parallel Layout Engine (February, 2012)
  4. Parallelism Tools Workshop (Intel, Santa Clara): Synthesizing a Parallel CSS Engine (August, 2011)
  5. HotPar 2011: Data Parallel Programming for Irregular Tree Computations (May, 2011)
  6. Qualcomm (Santa Clara): FTL: Fast Tree Language (October, 2010)
  7. IBM Research (Almaden): Parallel Browsers for the Mobile Era: Challenges, Solutions, and Opportunities (September, 2010)
  8. Qualcomm (San Diego): FTL: Fast Tree Language (poster, September, 2010)
  9. Parallel Bootcamp: A Parallel Browser: Lessons in Mobility and Irregularity (September, 2010)
  10. Stanford: A Parallel Browser for the Mobile Web (May, 2010)
  11. W2SP 2010: Secure Cooperative Sharing of JavaScript, Browser, and Physical Resources (May, 2010)
  12. Oakland 2010: ConScript: Specifying and Enforcing Fine-Grained Security Policies for JavaScript in the Browser (May, 2010)
  13. WWW 2010: Fast and Parallel Webpage Layout (April, 2010)
  14. Qualcomm: A Principled Approach to a Browser Layout Engine (April, 2010)
  15. Open House at Intel Labs: Parallel Webpage Processing (poster, March, 2010)
  16. OOPSLA 2009: Flapjax (October, 2009)
  17. Intel Developer Forum: Parallel Browser Algorithms (poster, September, 2009)
  18. Microsoft Research (Redmond): Fast Webpage Layout (September, 2009)
  19. Adobe: Fast Webpage Layout (September, 2009)
  20. #moz09 (Mozilla): It's the End of the Web as We Know It, and I Feel Fine (invited all-hands talk, April 30, 2009). contact for slides.
  21. CodeCon 2009: The Parallel Web Browser (biohackers and cypherpunks, April 19th, 2009). pdf slides
  22. HotPar 2009: The Parallel Web Browser (April, 2009)
  23. Samsung Research: Parallel Webpage Layout (invited seminar, February, 2009)
  24. Intel Languages Workshop: Update on the Parallel Browser Language (industry, 2008)
  25. BayFP: Flapjax & FIRE - Evolving a Reactive Web Language (industry, 2008). video!
  26. #moz08 (Mozilla): Implicitly parallel browser scripting (invited all-hands talk, 2008)
  27. Adobe: Mixed Imperative and Functional Reactive Programming in Flex and AS3 (research staff review + Flex team talk, 2007)
  28. Adobe: Flapjax: Functional Reactive Web Programming (Flex team talk, 2006)

Internal Talks

Controversial talks @ Berkeley not covered by the above:

Some External Talks by Better Presenters

Further Meyerovichs