Ian Arawjo

I am an Assistant Professor of Human-Computer Interaction at the University of Montréal in the Department of Computer Science and Operations Research (DIRO), where I am also affiliated with Mila. I lead the Montréal HCI group. In the recent past, I was a Postdoctoral Fellow at Harvard University, working with Professor Elena Glassman in the Harvard HCI group.

The Montréal HCI group is exploring three areas: augmenting programming practice with multi-modal AI, akin to notational programming; designing new AI-powered tools to support early-stage design processes; and developing interfaces to help end-users evaluate large language model (LLM) outputs, especially for code. As of fall 2024, the group is not currently recruiting new graduate students, although we are always open to collaborations. If you'd like to collaborate, peruse the research interests in my Letter to Prospective Graduate Students. Then, if you are still convinced, drop me a line at my email. If you are a stellar undergraduate, send me an email with a brief summary of who you are and what ongoing projects or domains you are interested in. There are opportunities to help out on existing projects, such as ChainForge, provided you have programming skill or experience analyzing data. Within UdeM, there are independent studies via IFT3150.

📩 When trying to get in touch, please see my communication policy. Although I do my best to reply to emails, I receive a large amount of emails each day and cannot answer all of them in a timely manner.

Biography and past work

I hold a Ph.D from Cornell University in Information Science, where I was advised by Professor Tapan Parikh. My dissertation work spanned the intersection of computer programming and culture, investigating programming as a social and cultural practice. I have experience applying a range of HCI methods, from ethnographic fieldwork, to archival research, to developing novel systems (used by thousands of people) and running usability studies. My first-authored papers have won awards at top HCI conferences, including at CHI, CSCW, and UIST.

Currently, I am the creator and lead developer of ChainForge, the first open-source visual programming environment for prompt engineering. I am developing ChainForge out in the open with colleagues at Harvard CS: Elena Glassman, Priyan Vaithilingam, and Martin Wattenberg. In the recent past, I invented notational programming, a paradigm of programming where you can handwrite notation (diagrams, writing, etc.) that inter-operates with traditional typewritten code.

Recent Events

Our recent pre-print, "Who Validates the Validators?", was cited as the direct motivation for new LLM-validator-oversight features in LangSmith and Autoblocks, LLMOps tools used by 100k+ users worldwide.
Priyan V. presented our work, "Imagining a Future of Designing with AI," at the 2024 ACM DIS Conference in Copenhagen, Denmark!
Participated in the ACM UIST Conference Program Committee Meeting (one of my submissions was conditionally accepted!)
Attended the 2024 ACM CHI Conference, whether I presented our work on ChainForge, attended the HEAL Workshop, and helped facilitate the first Dynamic Abstractions group meet-up
At the SEMLA LLMOps workshop at Polytechnique Montréal, I gave a talk on ChainForge ⛓️
At Concordia University, my alma mater, I gave a workshop on ChainForge to sociology students ⛓️
I gave a talk to Kenneth Huang's group at PennState University about ChainForge ⛓️
Attended PLATEAU 2024 in sunny Berkeley, CA ☀️
A pre-print of our paper on Antagonistic AI 😈 was released to the public, where it garnered some interest in the press.
I gave a talk at Microsoft Research Montréal about ChainForge ⛓️
Our paper on ChainForge was accepted to CHI 2024! 🎉
Our paper on "An AI-Resilient Text Rendering Technique for Reading Documents" was accepted to CHI 2024! 🎉

Some other things I did in the past:

Served as an intern in the Machine Learning Research group at Apple, under Megan Maher and David Koski
Supported and studied the Nairobi Play Project, a UNICEF Kenya CS education initiative founded by Ariam Mogos, supervised by Kentaro Toyama, Steve Jackson, and Tapan Parikh.
Developed a game to teach novices Javascript semantics, with Erik Andersen, François Guimbretière and Andrew Myers. This project launched a new, comprehension-first way to teach coding with minimal tutorials and was played by thousands of people.
Designed a psychological game theory model of read receipts, with Arpita Ghosh and David Easley.
Graduated from Concordia University in Montréal with a degree in Computation Arts and Computer Science. There, I was active in the Montréal Indie Games community and published a game featured by Apple on the App Store.

👋 Want to support my work, or just say thanks? See here.