Md Kowsher

Research Scientist, MetaPhD Student, University of Central Florida

I am a Research Scientist at Meta, where I work on multimodal and agentic LLMs for Facebook Trust & Safety, focusing on contextual understanding and reasoning. I am also a PhD student in Computer Science at the Center for Research in Computer Vision (CRCV), UCF, advised by Prof. Chen Chen.

My research focuses on efficient adaptation of large models — parameter-efficient fine-tuning, mixture-of-experts, and memory-efficient training — and on multimodal learning across text, image, and video.

Previously, I was an ML Research Intern at Apple (2026), a Multimodal LLM Intern at Meta (2025), a Research Scientist Intern at Nokia Bell Labs (2024), an Amazon Alexa Prize Researcher (2022–2023, 2nd place in SocialBot Grand Challenge 5), and an Applied Scientist Intern at Amazon (2022). I received my MSc in Computer Science from Stevens Institute of Technology, advised by Prof. Jia Xu.

Email CV Google Scholar GitHub Hugging Face LinkedIn

News

Oct 2026Joined Meta as a full-time Research Scientist, after my Meta internship in summer 2025.
Sep 2026Token Hypothesis accepted at NeurIPS 2026.
Aug 2026Monkey Jump accepted at EMNLP 2026 (Main) as an Oral.
May 2026Joined Apple (Cupertino) as an ML Research Intern for summer 2026.
May 2026LiME (Spotlight) and SliceFine accepted at ICML 2026.
2026Serving as an Area Chair for ACL Rolling Review (ARR).
Jan 2026FlowNIB accepted at ICLR 2026.
Oct 2025SliceFine, from my Meta internship, open-sourced by Facebook Research.
Jul 2025My first US patent application published: Infinite Reservoir Transformer.
May 2025Four papers accepted at ACL 2025 (2 Main, 1 Findings, 1 Workshop).
May 2025Joined Meta as a Multimodal LLM Intern.
Apr 2025Presented two papers and served as Volunteer Coordinator at NAACL 2025; received the NAACL Travel Award.
Nov 2024Propulsion accepted at COLING 2025 (Oral).
Oct 2024RoCoFT accepted at the NeurIPS 2024 FITML workshop (Oral).
Aug 2024Started my PhD at UCF with the ORCGS Doctoral Fellowship.
Jun 2024Joined Nokia Bell Labs as a Research Scientist Intern.
Sep 2023Our team won 2nd place in the Amazon Alexa Prize SocialBot Grand Challenge 5.

Publications

Highlighted papers are representative. Full list on Google Scholar.

2026

Where You Backpropagate Matters: Token Hypothesis for Memory-Efficient Fine-Tuning

Md Kowsher, Chen Chen

NeurIPS 2026

LiME: Lightweight Mixture of Experts for Efficient Multimodal Multi-task Learning

Md Kowsher, Haris Mansoor, Nusrat Jahan Prottasha, Ozlem Garibay, Victor Zhu, Zhengping Ji, Chen Chen

ICML 2026 Spotlight

paper project code dataset

SliceFine: The Universal Winning-Slice Hypothesis for Pretrained Networks

Md Kowsher, Ali O. Polat, Prasanth Murali, Ehsan Mohammady Ardehaly, Chen Chen

ICML 2026

paper project code

Monkey Jump: MoE-Style PEFT for Efficient Multi-Task Learning

Nusrat Jahan Prottasha, Md Kowsher, Chun-Nam Yu, Chen Chen, Ozlem Garibay

EMNLP 2026 (Main) Oral

paper project code

FlowNIB: An Information Bottleneck Analysis of Bidirectional vs. Unidirectional Language Models

Md Kowsher, Nusrat Jahan Prottasha, Shiyun Xu, Shetu Mohanto, Chen Chen, Niloofar Yousefi, Ozlem Garibay

ICLR 2026

paper code

2025

Predicting Through Generation: Why Generation Is Better for Prediction

Md Kowsher, Nusrat Jahan Prottasha, Prakash Bhat, Chun-Nam Yu, Mojtaba Soltanalian, Ivan Garibay, Ozlem Garibay, Chen Chen, Niloofar Yousefi

ACL 2025 (Main)

paper code

RoCoFT: Efficient Finetuning of Large Language Models with Row-Column Updates

Md Kowsher, Tara Esmaeilbeig, Chun-Nam Yu, Chen Chen, Mojtaba Soltanalian, Niloofar Yousefi

ACL 2025 (Main); NeurIPS 2024 FITML Workshop Oral

paper code

TituLLMs: A Family of Bangla LLMs with Comprehensive Benchmarking

Shahriar Kabir Nahin, Rabindra Nath Nandi, Sagor Sarker, Quazi Sarwar Muhtaseem, Md Kowsher, Apu Chandraw Shill, Md Ibrahim, Mehadi Hasan Menon, Tareq Al Muntasir, Firoj Alam

ACL 2025 (Findings)

paper code

LLM-Mixer: Multiscale Mixing in LLMs for Time Series Forecasting

Md Kowsher, Md Shohanur Islam Sobuj, Nusrat Jahan Prottasha, E. Alejandro Alanis, Ozlem Ozmen Garibay, Niloofar Yousefi

ACL 2025 TRL Workshop

paper code

Does Self-Attention Need Separate Weights in Transformers?

Md Kowsher, Nusrat Jahan Prottasha, Chun-Nam Yu, Ozlem Ozmen Garibay, Niloofar Yousefi

NAACL 2025

paper model

BnTTS: Few-Shot Speaker Adaptation in Low-Resource Setting

Mohammad Jahid Ibna Basher, Md Kowsher, Md Saiful Islam, Rabindra Nath Nandi, Nusrat Jahan Prottasha, Mehadi Hasan Menon, Tareq Al Muntasir, Shammur Absar Chowdhury, Firoj Alam, Niloofar Yousefi, Ozlem Ozmen Garibay

NAACL 2025

paper code

Propulsion: Steering LLM with Tiny Fine-Tuning

Md Kowsher, Nusrat Jahan Prottasha, Prakash Bhat

COLING 2025 Oral

paper code

Infinite Reservoir Transformer

Jia Xu, Md Kowsher

US Patent Application 18/780,055 (Pub. No. US20250200362A1)

patent

2024

Parameter-Efficient Fine-Tuning of Large Language Models using Semantic Knowledge Tuning

Nusrat Jahan Prottasha, Asif Mahmud, Md Shohanur Islam Sobuj, Prakash Bhat, Md Kowsher, Niloofar Yousefi, Ozlem Ozmen Garibay

Scientific Reports (Nature Portfolio), 2024

code

Token Trails: Navigating Contextual Depths in Conversational AI with ChatLLM

Md Kowsher, R. Panditi, Nusrat Jahan Prottasha, Prakash Bhat, Anupam Kumar Bairagi, Mohammad Shamsul Arefin

NLDB 2024

paper

L-Tuning: Synchronized Label Tuning for Prompt and Prefix in Large Language Models

Md Kowsher, Md Shohanur Islam Sobuj, Asif Mahmud, Nusrat Jahan Prottasha, Prakash Bhat

ICLR 2024 Tiny Papers

paper code

2023

Contrastive Learning for Universal Zero-Shot NLI with Cross-Lingual Sentence Embeddings

Md Kowsher, Md Shohanur Islam Sobuj, Nusrat Jahan Prottasha, Mohammad Shamsul Arefin, Yasuhiko Morimoto

EMNLP 2023 Workshop (MRL)

paper

NAM: From Hybrid Dialogers to Neural Responders

João Luís Lins, Nikhil Reddy, Abdul Rafae Khan, Md Kowsher, Abhijeet Gusain, Yeshwanth Reddy, Xuting Tang, Preet Jhanglani, Nusrat Zahan, Mengjiao Zhang, Yu Yu, Darshil Shah, Jia Xu

Alexa Prize SocialBot Grand Challenge 5 Proceedings (2nd place)

paper

Impact Learning: A Learning Method from Feature's Impact and Competition

Nusrat Jahan Prottasha, Saydul Akbar Murad, Md Kowsher, Apurba Adhikary, Sujit Biswas, Anupam Kumar Bairagi

Journal of Computational Science (Elsevier)

paper code

2022

Bangla-BERT: Transformer-Based Efficient Model for Transfer Learning and Language Understanding

Md Kowsher, Abdullah As Sami, Nusrat Jahan Prottasha, Mohammad Shamsul Arefin, Pranab Kumar Dhar, Takeshi Koshiba

IEEE Access, 2022

paper model

CARAN: A Context-Aware Recency-Based Attention Network for Point-of-Interest Recommendation

Md Billal Hossain, Mohammad Shamsul Arefin, Iqbal H. Sarker, Md Kowsher, Pranab Kumar Dhar, Takeshi Koshiba

IEEE Access, 2022

paper

Experience

Meta — Research Scientist, Menlo Park, CA

Oct 2026 – Present

Multimodal and agentic LLMs for Facebook Trust & Safety, focusing on contextual understanding and reasoning.

Apple — ML Research Intern, Cupertino, CA

May 2026 – Aug 2026

LLMs for knowledge-graph understanding and generation; fine-tuning and evaluating KG models.

Meta — Multimodal LLM Intern, Menlo Park, CA

May 2025 – Aug 2025

Efficient training of multimodal LLMs; proposed the Universal Winning-Slice Hypothesis and SliceFine.

UCF, Center for Research in Computer Vision — Research Assistant

Aug 2024 – Present

Vision-and-language research with Prof. Chen Chen.

Nokia Bell Labs — Research Scientist Intern, NJ

Jun 2024 – Aug 2024

Methods to make language models more accurate and efficient.

Stevens Institute of Technology — Graduate Research Assistant

Aug 2021 – Jun 2024

Reservoir Transformer for NLP and long-horizon time series forecasting, with Prof. Jia Xu.

Amazon Alexa Prize — Researcher (Team NAM)

Nov 2022 – Sep 2023

LLM-based conversational agents and conversation-driven image generation; 2nd place in SocialBot Grand Challenge 5.

Amazon — Applied Scientist Intern, Seattle, WA

May 2022 – Aug 2022

Deep learning models for ML monitoring systems.

Hishab Ltd. — AI Scientist, Dhaka

Apr 2020 – Aug 2021

Led a team building conversational agents and ASR for Bangla.

Education

University of Central Florida — PhD, Computer Science

2024 – Present

Multimodal learning, computer vision and language. Advisor: Prof. Chen Chen. ORCGS Doctoral Fellow.

Stevens Institute of Technology — MSc, Computer Science

2021 – 2023

NLP, LLMs and deep learning. Advisor: Prof. Jia Xu.

Noakhali Science and Technology University — BSc, Applied Mathematics

2015 – 2019

Honors & Awards

NAACL Travel Award, 2025
ORCGS Doctoral Fellowship, University of Central Florida, 2024
2nd Place, Amazon Alexa Prize SocialBot Grand Challenge 5, 2023
Full Member, Sigma Xi, The Scientific Research Honor Society, 2023
Graduate Fellowship, Stevens Institute of Technology, 2021
Research Excellence Award, Global Innovation & Excellence Awards, 2021
Best Paper Award, Int. Conf. on Cyber Security and Computer Science (Springer), 2020
Top Winning Award, Coronathon-19 Hackathon, Bangladesh, 2020
Best Paper Award, Int. Conf. on Computer, Communication, Chemical, Materials & Electronic Engineering (IEEE), 2019
Champion, Robi r-ventures 2.0, Robi Axiata Limited, 2019
National BASIS ICT Award, 2019

Service & Teaching

Area Chair: ACL Rolling Review (ARR), 2026
Reviewer: NeurIPS 2026, ICML 2026, ICLR 2024–2025, ARR 2023–2025, IJCNN 2024, ICDM 2023
Volunteer Coordinator: NAACL 2025
Teaching Assistant, Stevens Institute of Technology: CS 541 Artificial Intelligence (Spring 2024); CS 583 Deep Learning (Fall 2022 – Fall 2023); CS 800 Special Research Problem (Spring 2023)