Skip to Content

Daffodil Global · Research Division

Advancing Artificial Intelligence for a Smarter World

The Daffodil Global AI Research Lab pioneers breakthroughs in machine learning, computer vision, and natural language processing — translating cutting-edge science into real-world solutions.

98.4%Model Accuracy
240+Research Papers
42Active Projects
12+
Years of Research
85+
Research Scientists
240+
Published Papers
36+
Global Partnerships

Building AI That Understands, Reasons & Acts

At the Daffodil Global AI Research Lab, we believe that transformative AI should be responsible, explainable, and aligned with human values. Our interdisciplinary teams collaborate across continents to solve the hardest problems in modern AI.

From foundational model architecture to applied AI deployment, every project is anchored in scientific rigor and an unwavering commitment to ethical innovation.

Learn About Our Vision
🧠
Deep Learning
👁
Computer Vision
💬
NLP
🤖
Robotics
📊
Data Science
🔒
AI Safety
15+ Countries
Reached

Our Core Research Domains

Six interconnected domains where Daffodil Global AI Research is driving the frontier forward.

🧠

Deep Learning & Neural Architectures

Designing next-generation model architectures — transformers, graph neural networks, and sparse mixture-of-experts — that push performance boundaries.

Transformers GNN Sparse Models
👁

Computer Vision & Perception

Real-time object detection, 3D scene understanding, and multimodal vision systems for autonomous and industrial applications.

Object Detection 3D Reconstruction Multimodal
💬

Natural Language Processing

Large language models, cross-lingual understanding, sentiment analysis, and generative AI systems that communicate naturally and accurately.

LLMs Cross-lingual Generative AI
🤖

Robotics & Autonomous Systems

Integrating perception, planning, and control to build robots that adapt to dynamic, unstructured real-world environments.

Motion Planning Sim-to-Real RL
📊

Data Science & Analytics

Scalable data pipelines, causal inference, and explainable machine learning methods that make sense of massive, heterogeneous datasets.

Causal AI XAI MLOps
🔒

AI Safety & Ethics

Developing alignment frameworks, robustness techniques, and governance protocols to ensure AI systems remain safe, fair, and accountable.

Alignment Fairness Governance

Meet Our Research Team

World-class scientists, engineers, and domain experts united by curiosity and a passion for responsible AI.

🧑‍💻

Dr. Arif Rahman

Lab Director · AI Architecture

PhD MIT · 15 yrs deep learning research · 80+ publications

👩‍🔬

Dr. Priya Sharma

Lead Researcher · NLP

PhD Stanford · multilingual LLMs · cross-lingual transfer

👨‍🏫

Prof. Wei Zhang

Senior Scientist · Computer Vision

PhD CMU · 3D scene understanding · real-time perception

👩‍💼

Dr. Leila Karimi

Head of AI Safety

Oxford DPhil · alignment · robustness · AI governance

Selected Publications

Peer-reviewed work published at NeurIPS, ICML, CVPR, ACL, and other top-tier venues.

2024

Sparse Mixture-of-Experts with Adaptive Routing for Efficient LLM Inference

Rahman A., Zhang W., Sharma P. et al.

NeurIPS 2024
2024

Cross-Lingual Zero-Shot Transfer via Semantic Alignment Pretraining

Sharma P., Karimi L. et al.

ACL 2024
2023

Real-Time 4D Panoptic Segmentation for Autonomous Driving

Zhang W., Rahman A. et al.

CVPR 2023
2023

Scalable Alignment via Reward Model Ensembles and Debate

Karimi L., Rahman A. et al.

ICML 2023
View All Publications →

Trusted by & in partnership with

NVIDIA
Google DeepMind
MIT CSAIL
Microsoft Research
BUET
AWS AI

Ready to Shape the Future of AI?

Whether you're a researcher, engineer, or industry partner, we'd love to collaborate. Join the Daffodil Global AI Research Lab and help build intelligence that matters.