Bayan Bruss

VP of Applied AI Research @CapitalOne | Adjunct @Georgetown | Simple baselines, practical implementations.

[Tom Goldstein @tomgoldsteincs]

Our new guardian model lets you create LLM guardrails using natural text. This little 8B model efficiently checks in real time whether chatbots comply with bespoke moderation policies.

It's not often that academics beats industry models, but DynaGuard stacks up well!

[Monte Hoover @MonteBHoover]

There is still a lot of brittleness in getting guardian models to incorporate custom policies, but we think this is a step in the right direction. Try out DynaGuard in this interactive demo (and give us feedback to improve it!): https://t.co/ZArEreMkpK

[Monte Hoover @MonteBHoover]

Paper: https://t.co/oPWOZstRUQ
Models: https://t.co/DfcxGYpS7R
Dataset: https://t.co/HUKFLiNqsj

This was a collaborative effort with
@neeljain1717, @k_saifullaah, @taruschirag, @vatsalbaherwani, Joseph Vincent, Melissa Kazemi Rad, @cbbruss, @PandaAshwinee, @tomgoldsteincs

[Monte Hoover @MonteBHoover]

Guardrails with custom polices are hard for models trained on safety and harm-related datasets. But what if you trained a guardian model on arbitrary rules?
Introducing DynaGuard, a guardian model for custom policies: https://t.co/oPWOZstRUQ

[Furong Huang @furongh]

Your critic model is secretly a strong policy model. Stay tuned for a deep dive 🤩

[Bayan Bruss @cbbruss]

@chrmanning We had that for our flight back to Dulles this summer. It was as if they parked the plane at a different airport.

[Bayan Bruss @cbbruss]

@scaling01 Why are they even measuring it in pages. Tax_propmt_final_final_v2.docx

[Irina Rish @irinarish]

I am looking for a postdoc to lead projects related to this collaboration, on scaling laws, emergence and interpretability in pre- and post-training & inference/reasoning, in multimodal foundation models (language, time series, tabular data etc). HPC experience is a plus.

[Bayan Bruss @cbbruss]

@KezhiKong Awesome. Congrats Kezhi

[Bayan Bruss @cbbruss]

@natolambert @interconnectsai I hope the title of this article is a Dr Spaceman reference

[Bayan Bruss @cbbruss]

@AlexGDimakis I would take it a step further. Humans are really good at learning from non verbal social cues. A look that implies disappointment, excitement, frustration can be a profound reward signal in many situations.

[Epoch AI @EpochAIResearch]

The past 5 years have seen big successes in language, image and video generation, but relatively limited success in robotic manipulation. Why don’t we have laundry robots in every house?

One thing seems clear: training compute is not the blocker. 🧵

[Bayan Bruss @cbbruss]

If you share this belief, I’ve got some cool work for you to do

[will brown @willccbb]

i'm increasingly convinced that "transformative ai" is going to look like an abundance of specialized models for everything from drug design to weather sims to robotics to supply chains, not one agent to rule them all. we're going to need a lot more ai researchers

[Bayan Bruss @cbbruss]

@jxmnop We’re largely taking an interpreter approach to prompt optimizations. I wonder what a compiler approach looks like.

[Bayan Bruss @cbbruss]

@scaling01 Three years is the perfect prediction horizon for anything you want. It’s just close enough that people feel like it’s going to happen soon and just far enough that if the deadline passes no one will remember you were wrong.

[Bayan Bruss @cbbruss]

@DimitrisPapail Maybe we’ll invent gyms to simulate the natural use of our brains

[Bayan Bruss @cbbruss]

@ziv_ravid Denoise

[Bayan Bruss @cbbruss]

We have a mantra on the team “your data is trying to kill you”

[Taco Cohen @TacoCohen]

What I look for when hiring? EXTREME PARANOIA about code and data

[Bayan Bruss @cbbruss]

@abeirami @DimitrisPapail < 3 months for the greatest test taking technology invented to solve p6

[Mark Ibrahim @ICLR 2026 @marksibrahim]

Open-weights for our Llip multimodal vision-language model led by @lavoiems are public!

LLIP proposes new pre-training objective to capture the many ways to describe an image leading to strong performance across a suite of 22-zero shot benchmarks.

https://t.co/Tr354Kfcno