ML PhD student @GeorgiaTech, Math BS @USC. Deep learning theory, generalization, robustness.
Joined December 2019
- Tweets 914
- Following 699
- Followers 893
- Likes 3,237
63 Photos and videos
Pinned Tweet
1 May 2025
Excited to present at the first #AISTATS2025 poster session on May 3!
Ever wondered how LLMs can generalize to new tasks in-context despite only training on token completion? We formalize this phenomenon as "task shift" and investigate a linear version: arxiv.org/abs/2502.13285
1
2
23
2,596
May 24
The updated version of this paper has been accepted at @TmlrOrg šØš Very excited about implications of our results for SOTA robustness algorithms & understanding spurious correlations more generally. Journal version link: openreview.net/pdf?id=h81ztbā¦
23 Apr 2025
Heading to #ICLR2025 to present our SCSL workshop paper on understanding how last-layer retraining methods mitigate spurious correlations! openreview.net/pdf?id=B2W51aā¦
Stop by on Monday, April 28 to chat and learn more š
1
3
11
1,818
Mar 22
CramƩr on Lindeberg: "When he was reproached for not being sufficiently active in his scientific work, he said 'Well, I am really a farmer.' And if somebody happened to say that his farm was not properly cultivated, his answer was 'Of course my real job is to be a professor.'"
2
261
Tyler LaBonte retweeted
Multimodal reasoning with Phi-4-reasoning-vision, new work on scaling LLM inference, benchmarking AI agents in network operations, cinematic video generation, adaptive evaluation for LLMs, and using AI to improve individual and population health. msft.it/6013QiQgx
3
8
32
11,673
Mar 5
Our Phi-4-reasoning-vision-15B technical report is now available on arxiv:
arxiv.org/abs/2603.03975
1
5
408
Mar 4
Some nice coverage on our new model release, highlighting our hybrid approach to multimodal reasoning š
Mar 4
Microsoft built Phi-4-reasoning-vision-15B to know when to think ā and when thinking is a waste of time
venturebeat.com/ai/microsoftā¦
3
408
Mar 4
It's been the privilege of my career to help build the newest Phi series model from @MSFTResearch!
Phi-4-reasoning-vision-15B is open-weight & competitive on perf with 10X less compute/tokens.
Read the blog for math and CUA case studies, hybrid reasoning, data insights, & more!
Vision-language models improve multimodal systems, but can make them slower, costlier, and harder to deploy. Learn how Phi-4-reasoning-vision-15B, a compact and fast multimodal reasoning model, blends strengths of different methods while reducing their limits: msft.it/6014Q5X0u
ALT White line icons against a blue-green gradient background form an architecture flow chart. In the middle of the chart is a three-by-three matrix of circles and lines within a round-edge square. Above the matrix, three icons in a row ā an equation, a person using a desktop, and a head with gears flow by dotted lines to the matrix. To the left of the matrix is an icon representing a stack of files with an arrow pointing to the matrix. To the right of the matrix is a graph with a double headed arrow pointing to the matrix and to itself. Below the matrix is an icon representing a document. A dotted line arrow connects this graph to the matrix, showing the direction flowing from the matrix to the document. To the right of the document icon is an hourglass icon and three list icons with a dotted line connecting the hourglass to the lists.
10
887
Jan 14
Over the holidays, I stress-tested the AI coding hype by doing something concrete: I built a college football simulator game from scratch to see if agents actually deliver. Hereās what I learned š
2
1
185
Jan 14
Misc takeaways:
ā¢ Copilot GitHub was far more useful than I expected
ā¢ Keeping code style consistent across humans agents is painful
ā¢ Overall: Claude was best for agentic coding; Gemini best for interactive pair-programming
2
134
Jan 14
Finally, thanks to @Kangwook_Lee's "Tenure Track Simulator" post for inspiring me to make the game public and write this up!
1
92
11 Dec 2025
Nice work from GT colleagues about how next-token prediction naturally captures long-range structural dependencies!
10 Dec 2025
(1/6) Why does next-token prediction work so well, even for long text? š¤
Check out āProvable Long-Range Benefits of Next-Token Predictionā.
A rigorous explanation for LLMās long-range coherence/reasoning.
Joint work with Santosh Vempalaš arXiv: arxiv.org/abs/2512.07818
2
298
25 Nov 2025
Fara has been one of the most exciting projects to watch evolve @MSFTResearch over the last few months. From my perspective, Fara is a real advance towards natively multimodal computer-use agents (e.g., no accessibility trees). Congrats to Corby and the team on the release!
24 Nov 2025
Microsoft just dropped Fara-7B, its first on device AI Agent that can use your computer just like you would: it clicks, types, fills out forms and completed tasks just by āseeingā the screen. Itās best-in-class in terms of accuracy and cost from yours truly at Microsoft AI Frontiers and you can use it today
2
274
20 May 2025
Returning to Building 99 for my second internship @MSFTResearch working on multimodal reasoning. Come say hi!
1
88
5,704
6 May 2025
That's a wrap on ICLR/AISTATS! It was a wonderful experience to have deep research discussions in a part of the world I had never been to. Thanks to everyone who stopped by to chat or even just say hi š
1
17
1,180
5 May 2025
See you this afternoon (May 5) at Poster 42 in Hall A-E!
1 May 2025
Excited to present at the first #AISTATS2025 poster session on May 3!
Ever wondered how LLMs can generalize to new tasks in-context despite only training on token completion? We formalize this phenomenon as "task shift" and investigate a linear version: arxiv.org/abs/2502.13285
369