Dhruv Krishan

engineering @ DevRev | ucla’25 math/cs

[Dhruv Krishan @dkrishan8]

@skeptrune @morphllm yes

[Dhruv Krishan @dkrishan8]

@thdxr What does being “good” at inference mean?

[Dhruv Krishan @dkrishan8]

@TheIshanGoswami If you didn’t have the convinction to join them back then, you don’t deserve the money!

Also don’t see an offer here 😂

[Dhruv Krishan @dkrishan8]

@nateobrienn worst way to invest

[Dhruv Krishan @dkrishan8]

@benln @spacexai @cursor_ai Separately, if anyone is looking for a diwali party in SF

https://t.co/xoFuvczYPH

[Simar Singh @hellosimar]

Hosting a private Diwali evening with A.R. Rahman in SF.

- Only 50 people in the room. Curated group of founders, investors and artists.

- I'll be moderating a panel discussion with ARR and @krandiash (founder, @cartesia) on 'future of creativity in the AI age'

- A surprise release by Rahman and @rajavignesh (cofounders of Secret Mountain)

Diwali can't get bigger than this! Super excited.

[Dhruv Krishan @dkrishan8]

@jjkaplowitz No

[Simar Singh @hellosimar]

Hosting a private Diwali evening with A.R. Rahman in SF.

- Only 50 people in the room. Curated group of founders, investors and artists.

- I'll be moderating a panel discussion with ARR and @krandiash (founder, @cartesia) on 'future of creativity in the AI age'

- A surprise release by Rahman and @rajavignesh (cofounders of Secret Mountain)

Diwali can't get bigger than this! Super excited.

[Dhruv Krishan @dkrishan8]

@utk7arsh Excited to read, what use cases is this relevant for?

[Utkarsh @utk7arsh]

🌟 Excited to share that our paper, RoboReason, has been accepted to EMNLP 2026!

🤖 We study how well Vision-Language Models (VLMs) perform on spatiotemporal ordering tasks that use real-world robotics images.

Our key findings:
 🔹 Large VLMs often underperform previously reported results when evaluated on spatiotemporal ordering tasks using real-world robotics images.
 🔹 Overlaying specialized computer vision model outputs and increasing reasoning effort generally help improve large VLM performance. 
 🔹 Finetuning smaller models using GRPO and SFT on examples generated by specialized computer vision models often helps them achieve comparable results to much larger VLMs.

📍 Presenting with @sagnikanupam at @emnlpmeeting in Budapest 🇭🇺 tomorrow!

🗓️ Poster Session C | 9:00–10:30 AM
#EMNLP2026 #VLMs #LLMs #Robotics #AI

[Dhruv Krishan @dkrishan8]

@kartikeyjandial Wonder what went wrong with the agent

[Dhruv Krishan @dkrishan8]

Atleast make it lie @alexandr_wang 😂

[Dhruv Krishan @dkrishan8]

@bettersayAJ @neatlogs So excited

[Dhruv Krishan @dkrishan8]

@reachjalil @daleverett @Dale Add me to this as well

[Dhruv Krishan @dkrishan8]

@SoniaBaschez @ghostai

[Dhruv Krishan @dkrishan8]

@bettersayAJ @neatlogs Yes

[Kartikey Jandial @kartikeyjandial]

hired greg as @neatlogs swe😭

[Dhruv Krishan @dkrishan8]

@reachjalil B2C, obviously

[Dhruv Krishan @dkrishan8]

@ClarkDonovan7 Dropped an email, would love to chat as well!

[Dhruv Krishan @dkrishan8]

@alextaubman Hey @alextaubman, love what you guys are doing. Am an AI engineer, and an angel investor, would love to chat about the technology role