The future begins with a question.

Our applied research lab is a place for scientific curiosity. We explore ambitious questions, test ideas carefully and look for the bridge between what is possible and what is useful.

Explore openly. Experiment carefully. Build purposefully.

Our agenda focuses on applied questions at the intersection of AI, science and human needs. We value a well-designed experiment, a reproducible method and an honest account of what the evidence can tell us.

Research directions.

Explore how machine learning and generative systems can support useful tasks, with attention to reliability, evaluation and human judgment.
Language and multimodal systems, knowledge retrieval, model behavior and evaluation
Investigate language technologies that account for local context, including Bangla and multilingual information access.
Bangla language tools, multilingual interfaces, data quality and representation
Use computation, simulation and analysis to investigate scientific questions and develop reusable research tools.
Modeling and simulation, experiment tooling, scientific visualization
Understand how people interact with intelligent tools, and design experiences that make their capabilities and limits clearer.
Human-in-the-loop workflows, explainable interfaces, usability and accessibility
Explore methods for turning imperfect, fragmented information into useful, transparent decision support.
Data preparation, statistical analysis, uncertainty communication
Consider privacy, permissions, misuse, bias and reproducibility as part of the research process, from question to prototype.
Evaluation practices, data stewardship, transparent reporting

Individual studies begin with a defined problem, available resources and an evaluation plan.

Evidence is the way forward.

A promising demo is the beginning of an investigation. We want to understand when an idea works, where it fails and why.

1
Frame the question
Define a specific problem, map relevant work and write down assumptions before choosing a method.
2
Design the experiment
Identify suitable data, establish baselines and decide what useful evidence would look like.
3
Test and document
Record the method, compare outcomes and examine failure cases alongside positive results.
4
Share and translate
Explain findings and limitations clearly. Where appropriate, move the learning into a prototype or practical tool.