Simra Shahid

Our paper just won Best Paper at @sdpworkshop, ACL!! A special thank you to my co-authors who helped shape this project, and especially to @Hoper_Tom, from whom I have learned a lot through this collaboration!

@lmathur_ @GoogleDeepMind Congratulations

started working @GoogleDeepMind as a student researcher, and I'm excited for the summer ahead!

@lateinteraction How about an interview or podcast style video: one person dives into the technical details (with code) while the other explains a real-world use case. This structure might be more practical..

I frankly think I’ve done a mediocre job at explaining the right way to approach DSPy as a paradigm and tool. Thinking of making fresh material. What would make most sense. Regular technical threads? Videos where I rant while tinkering on a VS Code notebook? Lecture-y videos?

@mdancho84 Haha I am never forgetting this analogy!

Type 1 and Type 2 errors are confusing. In 3 minutes, I'll demolish your confusion. Let's dive in. 🧵

@bindureddy Why pay for a report you can generate too?

The fastest way to get rich! > use AI and create a deep research report on any technical topic > offer the research report at 100x cheaper cost than some fancy analyst firm > sell report and make $$ Disrupt Gartner, Forester, and every other research firm 😂

This is going front and center on my desk!! Done > Perfect. The ‘shouldn’t this be done this way?’ crowd might lose sleep over this one 😂

@ashleyyjoelle I would recommend ‘The Courage to Be Disliked’. I wish I had read it earlier 😅 For something different, ‘The Anthropologist’ isn’t life-changing, but I love the author’s writing style!

looking to read more this year…what’s a book that actually changed your life? any genre, just something that stuck with you

@jonfroehlich Absolutely love this collection! I think it goes beyond HCI—making figures in research papers more accessible and engaging for all disciplines is important.

HCI systems papers need to tell visual stories. As we near the CHI deadline, here's a set of exemplary figures curated by my lab. Hope it helps inspire! Feel free to contribute some examples. docs.google.com/presentation/d…

@lateinteraction I've been thinking about this too! As someone in NLP/IR, I've been connecting with researchers in HCI, RL and Data Sys. The conversations get really interesting. It's fascinating how abstracting jargon helps to identify important problems.

The types of problems I've been thinking about over the past few years has taught me that NLP people, IR people, PL/SE people, HCI people, 'mainstream' ML people, RL people, data systems people, MLSys people, and now the new "LLM" people barely talk to each other.

@JandialSurgan @abhinav_java @_cagarwal Thankyouu!!

@its_sshahid @abhinav_java @_cagarwal Just read through it ! Amazing insights in this work Simra! Congratsss

Accepted at #NAACL2025 How can individuals and organizations prevent their public text data from being exploited by open and closed language models? 🛡️ The answer lies in information theory!

Great collaboration with co-authors @abhinav_java and @_cagarwal!!

Check out the paper, blog, and code. Try this out on your classification data! Paper: openreview.net/forum?id=5QRQd… Code:  github.com/AikyamLab/regt… Blog: chirag-agarwall.github.io/latest_researc…

We asked 28 industry professionals about prompt engineering practices for deployment. Most of them spend a lot of time correcting overlooked instructions, leading to inconsistent outputs through trial and error. @ishikaajoshii & I built a tool to streamline this.

📢Thrilled to introduce the #VirtualLab: a team of AI scientist agents (AI chemist, AI reviewer...). Virtual Lab is led by an AI professor w/ feedback from human scientist. The Lab created new nanobodies that we experimentally validated to bind to recent #covid variants🚀🧵

Meet Context Autopilot It learns like you, thinks like you, and uses tools like you. With SoTA context understanding, it's capable of most information work today. Watch it beat a team of industry experts:

This is the best paper written so far about the impact of AI on scientific discovery

At @COLM_conf this week, check out our paper (presented by @CarlEdwards_NLP) on LLM-based drug synergy hypothesis generation. Predicting synergistic combination therapies given specific tumor information, and generating/retrieving drugs optimized to synergize in the context of specific patients. With novel forms of in-context learning, graph-based context optimization, and drug design task formulation via retrieval.