Katrina Drozdov (Evtimova)
@stochasticdoggo.bsky.social
AI researcher | PhD from NYU Data Science | Bulgarian yogurt, prime numbers, and dogs bring me joy | she/her | https://kevtimova.github.io/
Pinned
9 Lessons I Learned while Doing RL Post-Training for LLMs
I recently had the chance to experiment with post-training techniques for large language models, a space that has become central to making LLMs useful and controllable in real-world applications. I us...
kevtimova.github.io
Finally dipped my toes into RL post-training. I trained a code generation LLM with GRPO using open-r1. Here are my 9 takeaways: kevtimova.github.io/posts/grpo/
Finally dipped my toes into RL post-training. I trained a code generation LLM with GRPO using open-r1. Here are my 9 takeaways: kevtimova.github.io/posts/grpo/
9 Lessons I Learned while Doing RL Post-Training for LLMs
I recently had the chance to experiment with post-training techniques for large language models, a space that has become central to making LLMs useful and controllable in real-world applications. I us...
kevtimova.github.io
July 8, 2025 at 6:29 PM
Finally dipped my toes into RL post-training. I trained a code generation LLM with GRPO using open-r1. Here are my 9 takeaways: kevtimova.github.io/posts/grpo/
I asked ChatGPT, Gemini, and Claude for a clever joke. They all gave me the same one. Either AI is merging into a hive mind… or humor has officially been solved mathematically!
March 6, 2025 at 2:41 AM
I asked ChatGPT, Gemini, and Claude for a clever joke. They all gave me the same one. Either AI is merging into a hive mind… or humor has officially been solved mathematically!
The principle of least effort, from psychology, describes how we favor efficiency over effort. It aligns with System 1 (fast, intuitive) vs. System 2 (slow, deliberate) reasoning. AI faces a similar challenge: knowing when to rely on heuristics vs. deeper reasoning.
January 29, 2025 at 3:17 AM
The principle of least effort, from psychology, describes how we favor efficiency over effort. It aligns with System 1 (fast, intuitive) vs. System 2 (slow, deliberate) reasoning. AI faces a similar challenge: knowing when to rely on heuristics vs. deeper reasoning.
Reposted by Katrina Drozdov (Evtimova)
Happy Holidays and a Joyful New Year!
Wall sculpture series from “Random Walks of Happiness”, 2024
Celebrating Art: the Process, the Experiment, the Media.
#randomwalksofhappiness #happynewyear #happyholidays #abstractart #art #experimentalart #drawing #sculpture #wire #foundobjectsart #december
Wall sculpture series from “Random Walks of Happiness”, 2024
Celebrating Art: the Process, the Experiment, the Media.
#randomwalksofhappiness #happynewyear #happyholidays #abstractart #art #experimentalart #drawing #sculpture #wire #foundobjectsart #december
December 20, 2024 at 11:48 PM
Happy Holidays and a Joyful New Year!
Wall sculpture series from “Random Walks of Happiness”, 2024
Celebrating Art: the Process, the Experiment, the Media.
#randomwalksofhappiness #happynewyear #happyholidays #abstractart #art #experimentalart #drawing #sculpture #wire #foundobjectsart #december
Wall sculpture series from “Random Walks of Happiness”, 2024
Celebrating Art: the Process, the Experiment, the Media.
#randomwalksofhappiness #happynewyear #happyholidays #abstractart #art #experimentalart #drawing #sculpture #wire #foundobjectsart #december
Reposted by Katrina Drozdov (Evtimova)
If you are into ML theory (RL or not) with a proven track record, and you are interested in an industry research position, PM me. Feel free to spread the word.
December 19, 2024 at 12:55 AM
If you are into ML theory (RL or not) with a proven track record, and you are interested in an industry research position, PM me. Feel free to spread the word.
Reposted by Katrina Drozdov (Evtimova)
I gave a talk on Compositional World Models at NeurIPS last week 🌐
The recording is now online: neurips.cc/virtual/2024... (for registered attendees; starts at 6:06:00)
Workshop: compositional-learning.github.io
The recording is now online: neurips.cc/virtual/2024... (for registered attendees; starts at 6:06:00)
Workshop: compositional-learning.github.io
December 19, 2024 at 1:57 AM
I gave a talk on Compositional World Models at NeurIPS last week 🌐
The recording is now online: neurips.cc/virtual/2024... (for registered attendees; starts at 6:06:00)
Workshop: compositional-learning.github.io
The recording is now online: neurips.cc/virtual/2024... (for registered attendees; starts at 6:06:00)
Workshop: compositional-learning.github.io
Reposted by Katrina Drozdov (Evtimova)
Just 10 days after o1's public debut, we’re thrilled to unveil the open-source version of the technique behind its success: scaling test-time compute
By giving models more "time to think," Llama 1B outperforms Llama 8B in math—beating a model 8x its size. The full recipe is open-source!
By giving models more "time to think," Llama 1B outperforms Llama 8B in math—beating a model 8x its size. The full recipe is open-source!
December 16, 2024 at 9:42 PM
Just 10 days after o1's public debut, we’re thrilled to unveil the open-source version of the technique behind its success: scaling test-time compute
By giving models more "time to think," Llama 1B outperforms Llama 8B in math—beating a model 8x its size. The full recipe is open-source!
By giving models more "time to think," Llama 1B outperforms Llama 8B in math—beating a model 8x its size. The full recipe is open-source!
Reposted by Katrina Drozdov (Evtimova)
Announcement from TMLR jmlr.org/tmlr/
"""
📣 Heads-up 📣 that TMLR will pause new submissions over the upcoming holiday period from December 2 2024 to January 6 2025 (midnight AoE on both dates). We will resume accepting new submissions on January 7, 2025. Happy Holidays!
"""
"""
📣 Heads-up 📣 that TMLR will pause new submissions over the upcoming holiday period from December 2 2024 to January 6 2025 (midnight AoE on both dates). We will resume accepting new submissions on January 7, 2025. Happy Holidays!
"""
Transactions on Machine Learning Research
jmlr.org
November 26, 2024 at 2:05 PM
Announcement from TMLR jmlr.org/tmlr/
"""
📣 Heads-up 📣 that TMLR will pause new submissions over the upcoming holiday period from December 2 2024 to January 6 2025 (midnight AoE on both dates). We will resume accepting new submissions on January 7, 2025. Happy Holidays!
"""
"""
📣 Heads-up 📣 that TMLR will pause new submissions over the upcoming holiday period from December 2 2024 to January 6 2025 (midnight AoE on both dates). We will resume accepting new submissions on January 7, 2025. Happy Holidays!
"""
Reposted by Katrina Drozdov (Evtimova)
Write math on 🦋 with UnicodeIt!
For example: θ ∈ ℝⁿ or pp̅ → μ⁺μ⁻
Use website or install system-wide in Linux, macOS, or windows
www.unicodeit.net
(Created several years ago with @svenkreiss.bsky.social)
For example: θ ∈ ℝⁿ or pp̅ → μ⁺μ⁻
Use website or install system-wide in Linux, macOS, or windows
www.unicodeit.net
(Created several years ago with @svenkreiss.bsky.social)
November 23, 2024 at 9:34 PM
Write math on 🦋 with UnicodeIt!
For example: θ ∈ ℝⁿ or pp̅ → μ⁺μ⁻
Use website or install system-wide in Linux, macOS, or windows
www.unicodeit.net
(Created several years ago with @svenkreiss.bsky.social)
For example: θ ∈ ℝⁿ or pp̅ → μ⁺μ⁻
Use website or install system-wide in Linux, macOS, or windows
www.unicodeit.net
(Created several years ago with @svenkreiss.bsky.social)
Reposted by Katrina Drozdov (Evtimova)
If you post a research paper with a picture of an animal, I will follow you. It's the law.
November 21, 2024 at 7:23 PM
If you post a research paper with a picture of an animal, I will follow you. It's the law.
Reposted by Katrina Drozdov (Evtimova)
some little bluesky tips 🦋
your blocks, likes, lists, and just about everything except chats are PUBLIC
you can pin custom feeds; i like quiet posters, best of follows, mutuals, mentions
if your chronological feed is overwhelming, you can make and pin make a personal list of "unmissable" people
your blocks, likes, lists, and just about everything except chats are PUBLIC
you can pin custom feeds; i like quiet posters, best of follows, mutuals, mentions
if your chronological feed is overwhelming, you can make and pin make a personal list of "unmissable" people
November 20, 2024 at 11:56 AM
some little bluesky tips 🦋
your blocks, likes, lists, and just about everything except chats are PUBLIC
you can pin custom feeds; i like quiet posters, best of follows, mutuals, mentions
if your chronological feed is overwhelming, you can make and pin make a personal list of "unmissable" people
your blocks, likes, lists, and just about everything except chats are PUBLIC
you can pin custom feeds; i like quiet posters, best of follows, mutuals, mentions
if your chronological feed is overwhelming, you can make and pin make a personal list of "unmissable" people
Reposted by Katrina Drozdov (Evtimova)
Mat is not on 🦋—posting on his behalf!
It's time to revisit common assumptions in IR! Embeddings have improved drastically, but mainstream IR evals have stagnated since MSMARCO + BEIR.
We ask: on private or tricky IR tasks, are rerankers better? Surely, reranking many docs is best?
It's time to revisit common assumptions in IR! Embeddings have improved drastically, but mainstream IR evals have stagnated since MSMARCO + BEIR.
We ask: on private or tricky IR tasks, are rerankers better? Surely, reranking many docs is best?
November 20, 2024 at 7:47 PM
Mat is not on 🦋—posting on his behalf!
It's time to revisit common assumptions in IR! Embeddings have improved drastically, but mainstream IR evals have stagnated since MSMARCO + BEIR.
We ask: on private or tricky IR tasks, are rerankers better? Surely, reranking many docs is best?
It's time to revisit common assumptions in IR! Embeddings have improved drastically, but mainstream IR evals have stagnated since MSMARCO + BEIR.
We ask: on private or tricky IR tasks, are rerankers better? Surely, reranking many docs is best?
Reposted by Katrina Drozdov (Evtimova)
How many documents should you retrieve when using a reranker? The answer might surprise you!
Check out the excellent work from our intern Mathew on this important retrieval question. 👏
Check out the excellent work from our intern Mathew on this important retrieval question. 👏
Mat is not on 🦋—posting on his behalf!
It's time to revisit common assumptions in IR! Embeddings have improved drastically, but mainstream IR evals have stagnated since MSMARCO + BEIR.
We ask: on private or tricky IR tasks, are rerankers better? Surely, reranking many docs is best?
It's time to revisit common assumptions in IR! Embeddings have improved drastically, but mainstream IR evals have stagnated since MSMARCO + BEIR.
We ask: on private or tricky IR tasks, are rerankers better? Surely, reranking many docs is best?
November 20, 2024 at 8:15 PM
How many documents should you retrieve when using a reranker? The answer might surprise you!
Check out the excellent work from our intern Mathew on this important retrieval question. 👏
Check out the excellent work from our intern Mathew on this important retrieval question. 👏
Reposted by Katrina Drozdov (Evtimova)
Google Scholar is twenty years (and one day) old today https://www.infodocket.com/2024/11/18/google-scholar-turns-20-20-things-you-didnt-know-about-scholar/
November 19, 2024 at 12:35 AM
Google Scholar is twenty years (and one day) old today https://www.infodocket.com/2024/11/18/google-scholar-turns-20-20-things-you-didnt-know-about-scholar/
Reposted by Katrina Drozdov (Evtimova)
A whiteboard from our Jan 2023 team retreat where I wrote down our goals, and the top one was “10k user signups.”
We’re growing by 10k users every 10-15 minutes right now.
We’re growing by 10k users every 10-15 minutes right now.
November 16, 2024 at 2:46 AM
A whiteboard from our Jan 2023 team retreat where I wrote down our goals, and the top one was “10k user signups.”
We’re growing by 10k users every 10-15 minutes right now.
We’re growing by 10k users every 10-15 minutes right now.
Reposted by Katrina Drozdov (Evtimova)
I'm making a list of AI for Science researchers on bluesky — let me know if I missed you / if you'd like to join!
go.bsky.app/AcP9Lix
go.bsky.app/AcP9Lix
November 10, 2024 at 12:11 AM
I'm making a list of AI for Science researchers on bluesky — let me know if I missed you / if you'd like to join!
go.bsky.app/AcP9Lix
go.bsky.app/AcP9Lix
Reposted by Katrina Drozdov (Evtimova)
New here? Interested in AI/ML? Check out these great starter packs!
AI: go.bsky.app/SipA7it
RL: go.bsky.app/3WPHcHg
Women in AI: go.bsky.app/LaGDpqg
NLP: go.bsky.app/SngwGeS
AI and news: go.bsky.app/5sFqVNS
You can also search all starter packs here: blueskydirectory.com/starter-pack...
AI: go.bsky.app/SipA7it
RL: go.bsky.app/3WPHcHg
Women in AI: go.bsky.app/LaGDpqg
NLP: go.bsky.app/SngwGeS
AI and news: go.bsky.app/5sFqVNS
You can also search all starter packs here: blueskydirectory.com/starter-pack...
November 9, 2024 at 9:13 AM
New here? Interested in AI/ML? Check out these great starter packs!
AI: go.bsky.app/SipA7it
RL: go.bsky.app/3WPHcHg
Women in AI: go.bsky.app/LaGDpqg
NLP: go.bsky.app/SngwGeS
AI and news: go.bsky.app/5sFqVNS
You can also search all starter packs here: blueskydirectory.com/starter-pack...
AI: go.bsky.app/SipA7it
RL: go.bsky.app/3WPHcHg
Women in AI: go.bsky.app/LaGDpqg
NLP: go.bsky.app/SngwGeS
AI and news: go.bsky.app/5sFqVNS
You can also search all starter packs here: blueskydirectory.com/starter-pack...
Reposted by Katrina Drozdov (Evtimova)
November 4, 2024 at 10:01 AM
You can read about my research on emergent communication with adaptive compute at inference time and its connection to OpenAI's o1 model here: nyudatascience.medium.com/from-academi...
#ai #machinelearning #research
#ai #machinelearning #research
From Academia to Industry: How a 2018 Paper Foreshadowed OpenAI’s Latest Innovation
How a 2018 paper by CDS researchers helped shape OpenAI’s latest innovation, the o1 model.
nyudatascience.medium.com
November 14, 2024 at 3:41 PM
You can read about my research on emergent communication with adaptive compute at inference time and its connection to OpenAI's o1 model here: nyudatascience.medium.com/from-academi...
#ai #machinelearning #research
#ai #machinelearning #research
Hello Bluesky! 👋 I'm an AI researcher with a PhD from NYU's Center for Data Science. My research focuses on representation learning for images and video, with an emphasis on self-supervised learning and regularization methods. Excited to connect and explore here! #ai #deeplearning #computervision
November 14, 2024 at 3:32 PM
Hello Bluesky! 👋 I'm an AI researcher with a PhD from NYU's Center for Data Science. My research focuses on representation learning for images and video, with an emphasis on self-supervised learning and regularization methods. Excited to connect and explore here! #ai #deeplearning #computervision