Ph.D. Candidate @ Colorado State University

Efficient AI/ML
for Science

Specializing in building efficient AI/ML models. My Research interests are Multimodal Large Language Models, Foundational AI Models, Domain-Adaptation, Knowledge Distillation, and Trustworthy AI through Science Guidance. I build AI architectures that bridge the gap between massive pre-trained models and data-scarce scientific applications.

Abdul Matin

Professional Profile

I am a PhD Candidate and my research Focus is on bridging the gap between rigorous academic theory and impactful deployment. My work on optimizing Foundational Models for data-scarce environments has achieved a Best Paper Award at IEEE DSAA 2025 and active research contributor to NSF/NIFA funding projects, while my industrial experience includes building high-performance LLM based deep learning pipelines. With 8+ years of experience in research, education, and mentoring, I am seeking a position to lead Reliable AI innovations and mentor the next generation of talent in a collaborative research environment.

8+ Years
AI/ML Experience
35+
Total Publications
400+
Students Taught
30+
Advisees/Mentees

Research Vision: Trustworthy Multimodal Intelligence

As AI transitions from static prediction to complex reasoning, my agenda pioneers Reliable Foundation Models for the engineering and natural sciences.

Science-Aware Cognitive Agents

The Challenge: LMMs often "hallucinate" in scientific contexts where factual accuracy is non-negotiable.
Research Plan: Physics-Aware Multimodal RAG. I will be working on advancing agents that ground reasoning in verified observational data and physical laws.

Architecture: Physics-Constrained Decoding
+ Text
Multimodal Query
Foundation Model
Constraints Applied
Physically Valid Answer

Calibrated Trust in Dynamic Systems

The Challenge: High-stakes fields like precision agriculture require more than accuracy; they demand calibrated confidence.
Research Plan: Decomposing predictive uncertainty into Epistemic (reducible model knowledge) and Aleatoric (irreducible data noise) components to robustly detect out-of-distribution shifts.

Workflow: Uncertainty Decomposition
Dynamic Input
Prediction
Aleatoric
Epistemic
Safe
Flag OOD
Most Recent Research Highlight Best Paper Award - IEEE DSAA 2025

HyperKD: Distilling Cross-Spectral Knowledge in Masked Autoencoders via Inverse Domain Shift with Spatial-Aware Masking and Specialized Loss

We developed a novel framework that allows generalized knowledge from a widely capable but low-dimensional input based teacher to guide a high-dimensional input based student, solving the critical "dimensional scaling" challenge in foundational models.

Knowledge Distillation Foundational Models AI for Science
HyperKD Model Diagram

Click to view the HyperKD Architecture

Professional Experience

Graduate Research Assistant

2022 - Present
Colorado State University, Fort Collins, CO
  • Conducting research on domain-adaptive and multimodal AI using foundation models, knowledge distillation, and self-supervised learning.
  • Designing scalable learning methods that mitigate data scarcity in agricultural science and related decision-support applications.

Applied Scientist Intern (DL | LLM)

2025
Amazon, Palo Alto, CA
  • Developed a novel LLM-powered search & recommender pipeline by fine-tuning models to enhance query understanding and product-level relevance.
  • Optimized multi-step query refinement model’s inference by engineering a feature-rich input pipeline, resulting in a 50% reduction in online GPU/CUDA inference latency.

Applied Scientist Intern (ML)

2024
Amazon, Bellevue, WA
  • Developed a Machine Learning based logistics optimization model projected to achieve annual savings of $15M through efficient route planning and resource allocation.
  • Collaborated with applied scientists, software engineers, and operations teams to translate business constraints into scalable ML solutions.

Graduate Teaching Assistant

2021-2022
Colorado State University, Fort Collins, CO
  • Delivered recitation lectures and labs for four semesters, covering Big Data (CS535, CS435) and Computer Organization (CS270).
  • Guided 100+ graduate and undergraduate students through machine learning projects, resolved distributed cluster issues, and contributed to course material development.

Lecturer & Researcher

2016-2020
RUET, Bangladesh
  • Taught 300+ undergraduate students across five batches, delivering core courses including Artificial Neural Networks, Data Structures, Database Systems, and Object-Oriented Programming (C++).
  • Supervised and mentored 30+ undergraduate students on research initiatives and final-year theses, guiding several teams to successfully publish their work in international conferences.

Education

Ph.D. in Computer Science

Colorado State University (CSU)

Fort Collins, Colorado, USA

Expected Graduation: 2026

Fellowship for exceptional research excellence

M.Sc. in Computer Science

Colorado State University (CSU)

Fort Collins, Colorado, USA

Graduated: 2022

Outstanding Graduate Student Fellowship

B.Sc. in Computer Science & Engineering

Rajshahi University of Engineering & Technology (RUET)

Rajshahi, Bangladesh

Graduated: 2016

Topper & Vice Chancellor Gold Medalist

Technical Proficiency

Deep Learning

Visual Language Model (CLIP, LLaVA, PaliGemma) Large Language Model (Qwen, Llama) Visual Transformer Masked Autoencoder Informer Graph Neural Network CapsNet ResNet GAN LSTM TinesNet GBM

Programming

Python (Expert) C++ Java Scala JavaScript R

ML Frameworks and Clouds

PyTorch TensorFlow Apache Spark Hadoop PySpark AWS Git

Awards & Honors

Research Excellence & Fellowships

Best Paper Award

IEEE DSAA 2025

Awarded by IEEE DSAA, a flagship conference in data science and analytics.

Robert B. France Fellowship

Colorado State University, 2024

Prestigious fellowship awarded annually to one Ph.D. student for exceptional research excellence.

Outstanding Graduate Student Fellowship

CSU Dept. of Computer Science, 2023

Recognition for top-performing graduate student in the department.

Academic Honors & Competitions

  • Vice Chancellor Gold Medal (2016)

    Highest academic honor awarded by RUET to the top-ranked B.Sc. graduate.

  • RUET Student of the Year (4x)

    Awarded annually for highest performance (2012, 2013, 2014, 2015).

  • Champion, Divisional Inter-University Programming Contest (2014)

    1st Place, Hosted by ICT Division, Rajshahi.

  • Government Scholarships (2008 & 2010)

    For outstanding results in Secondary & Higher Secondary Board examinations.

Selected Publications

AAAI 2026 (KGML-Bridge Program)

Knowledge-Guided Masked Autoencoder with Linear Spectral Mixing and Spectral-Angle–Aware Reconstruction

Integrates the Linear Spectral Mixing Model (LSMM) as a differentiable constraint within self-supervised learning to respect real-world spectral composition laws.

Physics-Guided AIFoundation Modelscientific applications of HSI
KARMA Model Diagram

Click to view architecture

AAAI 2025

Accounting for Spatial Variability with the Histogram of Oriented Gradients Based Masking Improves Performance of Masked Autoencoder over Hyperspectral Satellite Imagery

Introduced a novel masking strategy, HOGMAE, based on the Histogram of Oriented Gradients that incorporates rich information inherent within satellite images during the mask creation step.

Masked AutoencoderFoundation ModelHigh dimensional Imaging
HOGMAE Model Diagram

Click to view architecture

IEEE BigData 2025

DeepSalt: Bridging Laboratory and Satellite Spectra through Domain Adaptation and Knowledge Distillation for Large-Scale Soil Salinity Estimation

Introduces the Spectral Adaptation Unit (SAU) to project high-precision lab spectroscopy and noisy satellite observations into a unified latent representation.

Domain AdaptationKnowledge DistillationHyperspectral Imaging
DeepSalt Framework Diagram

Click to view architecture

ACM SIGSPATIAL 2025

Terramae: Learning spatial-spectral representations from hyperspectral earth observation data via adaptive masked autoencoders

Proposed effective Masking for self-supervised MAE and Enforces structural and spectral coherence via specialized loss functions (SSIM/SID) to learn physically meaningful features from Earth observation data.

Self-Supervised LearningFoundation ModelGeospatial AI
TerraMAE Architecture Diagram

Click to view architecture

IEEE BigData 2023

Discern: Knowledge Distillation for High Resolution Soil Moisture Estimation

Bridges sparsity and resolution by distilling coarse-resolution predictions from a large teacher into high-resolution outputs for a lightweight student anchored by sparse sensors.

Model CompressionKnowledge DistillationScientific ML
Discern Model Diagram

Click to view architecture

Google Scholar Profile

Funded Research Projects

Interdisciplinary research projects I contributed to as a graduate research assistant (* indicates my Doctoral Committee Members).

NIFA DSFAS 07/2024 – 06/2027

Enabling Effective Decision-Making In Dryland Farming For Arid And Semi-Arid Regions Using Field- scale Soil Moisture Content (SMC) Maps

Collaborators/PIs: Sangmi Lee Pallickara* (PI) with Allan Andales, Jeffrey Niemann*, and Shrideep Pallickara*.
$591,500 Grant Amount
NSF CPS 08/2023 – 07/2026

Medium: Making Every Drop Count: Accounting for Spatiotemporal Variability of Water Needs for Proactive Scheduling of Variable Rate Irrigation Systems.

Collaborators/PIs: Sangmi Lee Pallickara* (PI) with Allan A. Andales, Jeffrey Niemann*, and Shrideep Pallickara*.
$1,199,846 Grant Amount
FFAR / Noble Research 10/2021 – 09/2026

Metrics, Management, and Monitoring – An Investigation of Pasture and Rangeland Soil Health and Its Drives.

Collaborators/PIs: Michigan State Univ. (Jason Rowentree), CSU (F. Cotrufo, K. Paustian, S.L. Pallickara*), Noble Research Institute, Oregon State Univ.
$20 Million Total Scope

Mentorship & Advising

Beyond my teaching duties, I have provided direct mentorship to more than 30 undergraduate and graduate students, supporting their undergraduate theses, research endeavors, and software development projects.

Mentorship Directory

Listing research advisees and thesis students mentored across CSU and RUET.

Student Name Role Research / Mentorship Topic Outcome / Current Status
Rupasree Dey Graduate Mentee Research on DeepSalt: Domain Adaptation for Soil Salinity Published in IEEE BigData '25
Tanjim B. Faruk Graduate Mentee Research onTerraMAE: Hyperspectral Feature Learning Published in ACM SIGSPATIAL '25
Hafsa Binte Kibria Undergrad Advisee Research onCardiovascular Disease Prediction Models Faculty Member @ ECE, RUET
Omaer Faruq Undergrad Advisee Engineering Projects & Cloud Segmentation Faculty Member @ ECE, RUET
Hasinur Rahman Undergrad Advisee Data Structures & Algorithms Software Engineer @ Google
Tonmoy Hasan Undergrad Advisee Research on Sentiment Analysis & Text Classification Ph.D. Student @ UNC Charlotte
MD Rakibul Islam Undergrad Advisee Research on Deep Learning for Medical Imaging (CT/MRI) Ph.D. Student @ Mälardalen Univ
Ettilla Mohiuddin Eumi Undergrad Advisee Research on Bengali News Categorization (NLP) Ph.D. Student @ UNSW
Habib Ur Rahman Undergrad Advisee Undergraduate Project: Software Development & Machine Learning Ph.D. Student @ CSU
S. Sarker Undergrad Advisee Research on Precipitation Prediction using LSTM Published in IJCI 2022
I. Z. Tonu Undergrad Advisee Project and Research on Human Action Classification (CNN) Published in IEEE WIECON 2021
S. Islam Undergrad Advisee Research on Hand Gesture Recognition (HCI) Published in ICO 2021
M. O. F. Goni Undergrad Advisee Research on Chronic Kidney Disease Prediction Published in BIM 2021
M. R. I. Sarker Undergrad Advisee Research on Collaborative Recommendation Systems Published in ICTSD 2021
M. A. I. Siddique Undergrad Advisee Research on White Blood Cell Classification Published in IEEE ICECE 2020
M. F. Ahamed Undergrad Advisee Software Developement Project and Research Published in ICAICT 2020
5+ Years
Teaching Experience
400+
Students Taught
30+
UG Projects/Theses Supervised
15+
Publications From Mentees

Service & Leadership

Reviewer

Reviewer for high-quality international conferences and journals in Machine Learning domain.

Technical Chair

RUET ECE Fest & IUPC 2019. Oversaw technical operations for national events with 90+ universities.

Technical Committee Member

ICEEE 2017 & ICECTE (2016, 2019). Managed technical evaluation processes for IEEE-affiliated conferences.

Organizing Committee

RUET ECE Fest (Technocracy 2018). Coordinated logistics and student engagement for a major engineering festival.

Technical Support Team

NCPC 2015. Provided on-site support for system integrity during the National Collegiate Programming Contest.

Let's Collaborate

I am always open to discussing multimodal AI, scientific machine learning, or potential research collaborations. Whether you have a question about my papers or just want to say hi, my inbox is open.

Email Me Directly

abdul.matin@colostate.edu

Connect on Socials

This will open your default email client.