Aditya Arora

Reproducibility crisis in AI/ML? ICML 2026

Great initiative by @AalokDThakkar A regularly updated repository of grants, fellowships, and awards in the sciences for students, postdocs, and faculty. aalok-thakkar.github.io/grants Do check it out. Worth bookmarking!!

Announcing the Voice Arena STT Leaderboard, our human-verified benchmark testing how accurately speech-to-text models transcribe real, spontaneous conversation across 7 languages The leaderboard covers 🇺🇸 US English, 🇮🇳 Hindi, 🇧🇩 Bangla, 🇧🇷 Brazilian Portuguese, 🇻🇳 Vietnamese, 🇷🇴 Romanian and 🇸🇦 Arabic. Every model is evaluated on the same held-out corpus per language: real, unscripted phone calls between native speakers, recorded in-country on their own phones, with the noise, fillers and interruptions of natural conversation. Unlike public STT benchmarks, results are measured on fully proprietary audio - no model has seen a second of it in training. @Reson8Offical's resonant-1 leads the US English board at 4.34% WER, ahead of models from @Microsoft , @Google , @OpenAI and @Meta . @SarvamAI leads in Hindi and Bangla, while Microsoft takes #1 in Romanian and Brazilian Portuguese. Key elements of the Voice Arena STT Leaderboard: ➤ Real speech, real conditions: audio is drawn from unscripted phone calls between native speakers, recorded in-country on the speakers' own devices, preserving the background noise, fillers and interruptions of production traffic. ➤ 100% proprietary, held-out audio: the corpus is collected and owned by Voice Arena. No model has seen a second of it in training, so results cannot be inflated by dataset contamination. ➤ Human-verified ground truth: transcripts are produced by native transcribers through a six-stage pipeline, and only segments verified 3-of-3 survive into the final test set. ➤ Corpus-level WER with symmetric, language-aware normalization: the same normalization is applied to both references and hypotheses, so models are penalized for recognition errors, never for spelling choices. Key results for the Voice Arena STT Leaderboard: ➤ The most accurate US English model is not from a big lab: Reson8's resonant-1 leads at 4.34% WER, ahead of Microsoft, Google, OpenAI and Meta. ➤ Model choice alone is worth ~44%: on the US English board, #1 sits at 4.34% WER and #20 at 7.74%. Same audio, same test set. Picking the right model cuts nearly half of your transcription errors before you touch anything else in the pipeline. ➤ There is no global winner: the #1 model changes with the language, with Reson8 leading US English, Sarvam leading Hindi and Bangla, and Microsoft leading Romanian and Brazilian Portuguese. Teams building for more than one market will not be covered by a single vendor at the top of every board.

I am looking for PhD students to join my lab at Johns Hopkins University in Fall 2027. If you’re attending #ACL2026 in San Diego, feel free to contact me through the Whova app. I’ll be around today and tomorrow. @aclmeeting @jhuclsp @JohnsHopkins @HopkinsEngineer

I'm hiring for my team at GDM! If you’re passionate about engineering scalable agentic learning systems and doing impactful research, please apply here: goo.gle/4wsFoTU

Disappointed by the review received from the tutorial track in @emnlpmeeting and @aaclmeeting literal gatekeeping…

I’m hiring! Come join our team at Google DeepMind in London or Mountain View to work on Gemini agent post-training. We are looking for Research Scientists and Research Engineers interested in advancing the capabilities of AI agents. Please apply here: google.com/about/careers/…

Joining @openai next month! after seeing people's reaction to Alisa's post about her experience, I also wrote down some of the surprising things I wish I know before my research scientist job search: yongzx.github.io/blog/2026/06/2…

I owe it all to this Reddit post

Final version of my book (with a new title) Online Learning: A Modern Introduction Using Convex Optimization Especially proud of the Foreword by @NicoloCB! It'll be printed by Cambridge University Press. The end of 7 years of updates :) arxiv.org/pdf/1912.13213

I am making the materials for my undergrad Compiler course freely available: LigerLabs.org/compilers.html. There are currently 26 lecture modules, each consisting of a ~20 minute video, slides, in-class exercises, and take-home assignments. Support: NSF/SATC/EDU.

No one benefits from pages of dense math. Ultimately, only the loss function (and the corresponding code) matter. The paper could have been a blog post: leetarxiv.substack.com/p/rectified-fl…

1/ We think deep reasoning models need complex hierarchical loops, slow/fast memory tracks, or explicit scratchpads. They don't. A flat, single-loop Transformer can outperform them all. It just needs the right signal propagation. 🧵

“I started my process by watching all the lectures from Stanford’s Language Modeling from Scratch course, which is helpful for illustrating the breadth of topics I needed to learn and helped me organize many scattered concepts in my brain into one coherent picture of the field.”

this PhD student had 47 interviews and 4 offers before she was hired at OpenAI. she practiced with her “notes on LLMs” and math and they’re a goldmine. super concise and organic and shared to everyone for free. you can use her notes or her topic list to study on your own.

@CSProfKGD @akshitac8 Thank you @CSProfKGD for showing us around the beautiful city of Zurich 🇨🇭

You know your former students like you when they reach out to meet up ❤️ Thanks @imAArora (former MSc student) and @akshitac8 for a FUN morning in Zurich!

LLM community slowly rediscovering what we in vision found out over half a decade ago. MY SCHMIDHUBER MOMENT IS COMING! Source: S4L paper where i tuned the most sota 10% and 1% ImageNet baselines ever, by far. arxiv.org/abs/1905.03670

This may be a controversial take, but I think it needs to be said: the gap between computer vision research in academia and industry is widening with every conference. A huge fraction of @CVPR papers—especially those that boil down to "we tweaked/fine-tuned/RL'ed large-scale model X to improve on task Y"—will become obsolete with the next model release. That's not where academia creates lasting value. PIs should adapt much faster to this changing reality. Academia should focus on fundamentally new ideas, new problem formulations, explaining emergent phenomenology, or uncovering blind spots that industry can later solve with scale, compute, and data.

Great quote that every AI researcher should aspire to: "I do things, which work, rather than being novel, but not working." - @ducha_aiki

Just selected as ICML oral! Congrats and great appreciation to all collaborators!!