Senior Applied Scientist @Microsoft. NLP/IR Ph.D. from @umasscs.
Amherst, MA
Joined May 2017
- Tweets544
- Following1.2K
- Followers1.1K
- Likes8.6K
The 2nd workshop on CustomNLP (models customized to individuals, groups, specialized domains) is back at #ACL2026! Join us if you're around! Schedule 👇👇👇👇👇
🚨 Join us on July 3 for the 2nd CustomNLP Workshop at #ACL2026!
We're excited to share a fantastic lineup of invited talks & papers on personalizing, adapting, and customizing LLMs. Hope to see you there! 🎉
📅 customnlp4u.github.io/schedu…
👇 Meet our amazing invited speakers:
Sheshera Mysore retweeted
📣We are looking for reviewers for our #CustomNLP4U workshop at #ACL2026! If interested, please sign-up here 👉 forms.gle/5M2syqsUK4Cr7HAD7
Submit your work on customizing models to users, groups, domains to our workshop at ACL 2026 in San Diego! 🌊🌊
The deadline is 3rd April - short (4 pages) and long (8 pages) papers are welcome - its a nice opportunity to get feedback from a strong group of reviewers!🔬🔬
📣 Deadline is coming soon this Friday (April 3rd) 11:59pm AoE!
We’re looking forward to your work on customizable NLP! ⭐️
#NLProc #ACL #CustomizableNLP
Sheshera Mysore retweeted
⚠️New paper on LLM training w/ private datasets.
With RL, we can now use aggregated statistics from a private dataset to train LLM / agents, so that data owners don't have to share their raw content to model developer.
Paper: arxiv.org/abs/2602.18633
Sheshera Mysore retweeted
Fantastic efforts towards adapting LLMs to users, groups, communities, organizations, and beyond.
📣 We are BACK at #ACL2026 in San Diego☀️🌊
If you’re tackling problems related to customizable NLP, consider submitting your work to our 2nd workshop! 🙌
More details:
🔗 customnlp4u.github.io/
#ACL #NLProc #ACL2026NLP #customizableNLP @aclmeeting
The workshop on Customizable NLP is back at #ACL2026!
Making LLMs truly useful for the broadest set of users means adapting them to the workflows of individuals 👤, teams 👥, and organizations 🏢.
Submit your work (data 📊, models 🤖, implications 🌍) to us!
🗓️ Due: 3rd April.
📣 We are BACK at #ACL2026 in San Diego☀️🌊
If you’re tackling problems related to customizable NLP, consider submitting your work to our 2nd workshop! 🙌
More details:
🔗 customnlp4u.github.io/
#ACL #NLProc #ACL2026NLP #customizableNLP @aclmeeting
Sheshera Mysore retweeted
Call for Applications - The CIIR Internship Program - Summer 2026
The Center for Intelligent Information Retrieval (CIIR) at UMass Amherst is accepting research internship applications.
See the thread for more information.👇
Sheshera Mysore retweeted
i used the first version of @openreviewnet to submit one of total 47 papers to the very first @iclr_conf in 2013. since then, i've had a privilege to use @openreviewnet as an author, a reviewer, an area chair, a program chair and a workshop organizer. now i'm super happy to be a supporter of @openreviewnet as well. join me!
🚀We are hiring research interns at Microsoft! 🚀
Apply!: apply.careers.microsoft.com/…
We're a high energy and well resourced research team interested in: training models using in-the-wild interactions, RL environments, evaluating agents for long-horizon + multi-party tasks.
Sheshera Mysore retweeted
I'll be at #NeurIPS2025!
📢 We are hiring full time scientists, machine learning engineers, and research interns at the Office of Applied Research 📢
If you are into LLM agents and/or large-scale RL training and eval, I would love to connect!
Thrilled that our paper analyzing Microsoft Copilot for consumer and WildChat-1M logs received a Senior Area Chair highlight award at @emnlpmeeting!
I'm excited for work on "LLMs in the wild" and grateful for all the energy and interest we received at the #EMNLP2025!
🤔 We know what people are using LLMs for, but do we know how they collaborate with an LLM?
🔍 In a recent paper we answered this by analyzing multi-turn sessions in 21 million Microsoft Copilot for consumers and WildChat interaction logs: arxiv.org/abs/2505.16023
ALT Overview of the key questions, data, and idea driving our analysis.
I'll be at #EMNLP2025 in Suzhou, China presenting this work! Get in touch if you'll be around and are free to catch up or meet!
🤔 We know what people are using LLMs for, but do we know how they collaborate with an LLM?
🔍 In a recent paper we answered this by analyzing multi-turn sessions in 21 million Microsoft Copilot for consumers and WildChat interaction logs: arxiv.org/abs/2505.16023
ALT Overview of the key questions, data, and idea driving our analysis.
Sheshera Mysore retweeted
📣Hiring! Two opportunities
1. Research internship@MSR AI Interaction and Learning (Current PhD students)
2. Multiple positions in my lab @UTiSchool on LLM Personalization/Human-AI Alignment (Prospective PhD students)
Details in thread below👇
Thrilled that this work was accepted to #EMNLP2025!
It presents the first large scale analysis of users collaborating with LLMs in-the-wild! Also thrilled that this kind of NLP+HCI work received generous support (and feedback) from reviewers and landed in the top 15% of papers!
🤔 We know what people are using LLMs for, but do we know how they collaborate with an LLM?
🔍 In a recent paper we answered this by analyzing multi-turn sessions in 21 million Microsoft Copilot for consumers and WildChat interaction logs: arxiv.org/abs/2505.16023
ALT Overview of the key questions, data, and idea driving our analysis.
Sheshera Mysore retweeted
We (Microsoft's Office of Applied Research) have a few openings for exceptional Research Engineers / MLEs to train, test, and scale the next-gen enterprise agentic systems to millions of copilot users. Apply today if you are up for the task!
jobs.careers.microsoft.com/g…
Sheshera Mysore retweeted
Landed in Vienna for #ACL2025! We are hiring FTEs/Postdocs/Interns at Office of Applied Research to push the frontier of continuous model improvement for productivity, through RL*, inference time scaling, self reflection, memory etc. Available to chat this week w/ @mengtingwan.
Sheshera Mysore retweeted
Great to see multiple efforts digging into real chat data! This new paper looks at how people actually collaborate with AI when writing (e.g., revising, co-creating). A nice complement to our WildFeedback work on learning from natural user feedback.
arxiv.org/abs/2505.16023
🤔 We know what people are using LLMs for, but do we know how they collaborate with an LLM?
🔍 In a recent paper we answered this by analyzing multi-turn sessions in 21 million Microsoft Copilot for consumers and WildChat interaction logs: arxiv.org/abs/2505.16023
ALT Overview of the key questions, data, and idea driving our analysis.
🤔 We know what people are using LLMs for, but do we know how they collaborate with an LLM?
🔍 In a recent paper we answered this by analyzing multi-turn sessions in 21 million Microsoft Copilot for consumers and WildChat interaction logs: arxiv.org/abs/2505.16023
ALT Overview of the key questions, data, and idea driving our analysis.
🧭What this means for LLM alignment 🧭
Most excitingly, we found LLMs being used as partners in 🧠learning 🧠. Aligning LLMs for learning represents a big opportunity for interdisciplinary work and our paper describes this and other under-explored research to explore in future.
ALT Broad directions of future work implied by out findings.
1
1
376
This was a deeply fulfilling collaboration at the Office of Applied Research at Microsoft with @geekylildeb , @CaoHancheng , and Bahar Sarrafzadeh.
📄 Paper: arxiv.org/abs/2505.16023
📝 Blog (4-min read): prototypical-hai-collaborati…
💻 GitHub: github.com/microsoft/prototy…
3
322