DeCLaRe Lab is a good fit for people who enjoy building AI systems, asking careful scientific questions, and working across safety, trustworthiness, multimodality, AI for Science, efficiency, and embodied AI.

Why Join DeCLaRe Lab?

Project-centered research

Work on models, benchmarks, demos, and open-source systems that connect papers to working artifacts.

Collaborative culture

DeCLaRe is built around collaborative research, where students work closely with lab members, academic collaborators, and industry partners to shape ambitious AI projects.

Active research program

The lab is recruiting at NTU across the six core themes of safety, trustworthiness, multimodality, AI for Science, efficiency, and embodied AI.

Collaboration structure

How Students Collaborate in DeCLaRe

A star-schema view of coauthorship patterns in the lab's publication archive from 2019 onward: papers sit at the center, with author, collaborator, partner, and research-theme dimensions around them.

Papers analyzed 167

Publication archive, 2019 onward

Average authors per paper 5.6

Median: 5 authors

Non-PI DeCLaRe members per paper 1.5

Current members and alumni; median: 1

DeCLaRe type-level collaboration network A star-schema view of the publication archive, with papers at the center and author, collaborator, partner, and research-theme dimensions around it. 49.1% 84.4% 89.8% 98.0% 21.3% 70.1% Publicationarchive InternalDeCLaRecollaboration Initiatedin DeCLaRe Externalcollaborators Academicpartners Industrypartners Multi-themepapers

What this means for students

Students can work across themes and often collaborate with other DeCLaRe-affiliated authors and external research partners. The graph is intentionally type-level: it emphasizes the lab's collaborative structure rather than ranking individuals.

How to read the numbers

Percentages summarize papers from 2019 onward; DeCLaRe members include current members and alumni. Internal collaboration excludes the PI.

Current Openings

PhD Students

For applicants interested in safety, trustworthy AI, multimodal systems, AI for Science, efficient learning, or embodied AI.

Postdoctoral Researchers

For researchers with a strong publication record and a clear agenda connected to the lab's themes.

Research Interns and Visiting Researchers

For students and collaborators who can contribute to focused projects, benchmarks, demos, or open-source releases.

Expected Backgrounds

Strong candidates often have experience in one or more of the following:

Machine learning Natural language processing Computer vision Audio and speech Robotics or embodied AI AI for Science Efficient learning Human-computer interaction AI safety and alignment Open-source engineering

How to Apply

Please review the lab's research slides, browse recent publications, and identify the research directions that genuinely connect with your interests.

Send a concise email to soujanya [dot] poria [at] ntu [dot] edu [dot] sg with your CV, transcript if applicable, representative work, and a short note on the projects you would like to pursue.

FAQ

Should I email before applying formally?

Yes, a short, specific email is helpful. Mention the lab theme or paper that connects to your interests.

Do I need prior publications?

Publications help, but strong engineering ability, research taste, and evidence of persistence also matter.

Can I propose a new direction?

Yes. The best proposals still connect clearly to DeCLaRe Lab's core agenda.