Social Computing × AI × Governance

I am a PhD student at the UMich working with Farnaz Jahanbakhsh. I build AI-powered social computing systems that give users control over their digital experiences—from personalized content moderation to focused browsing.

When I encounter a problem, I love to design and code solutions (see here, also here).

How can we design and build systems that empower users to shape their digital experiences?

Research

Personalized Content Moderation Through AI

Platforms define harm universally and respond with suppression. What if users could define harm themselves? We built DIY-MOD, a system with optimized LVLM pipelines that transforms user-defined harmful elements within content in real-time instead of removing the entire post. (CHI 2026, under review)

Current work broadening the scope of AI-powered user safety and autonomy:

Agentic AI for Online Safety: How can autonomous systems protect users across the entire web, beyond specific platforms?

On-Device Intelligence with WebGPU: Can we democratize AI safety by eliminating cloud dependencies? I’m building hybrid architectures where sensitive processing happens locally while leveraging cloud models strategically—making personalized safety both private and affordable.

Governance for User-Centered Platforms

I think systems alone don’t create change, we also need regulatory frameworks that enable user agency. My Consumer Moderation Rights framework (FAccT 2026, in prep) connects user autonomy to existing consumer protection laws, providing actionable pathways for platform accountability.

LLMs for Scaling Education

I believe LLMs’ highest impact lies in democratizing quality education:

– Current: In collaboration with UMich Physical Therapy, we built LLM systems that scale expert feedback for clinical documentation and learning (AAAI IAAI 2026, under review)

– Previous: In 2023-24, I worked on developing one of the first LLM-based personalized tutoring systems with virtual tutors called KALEEM. (demo)

Selected Publications

• What If Moderation Didn’t Mean Suppression? A Case for Personalized Content Transformation
  Rayhan Rashed, Farnaz Jahanbakhsh | CHI 2026 (Under Review)
  📄 PDF | Slides | Code

• Consumer Moderation Rights: Fixing Social-Media “Defective Design”
  Rayhan Rashed | FAccT 2026 (Manuscript Under Prep)
  📄 PDF

• Learning Together: Improving Documentation in a Pro Bono Physical Therapy Clinic
  Rayhan Rashed et al. | IAAI 2026 (Under Review)
  📄 PDF

• What a Large Scale Study Tells Us About the Digital Divide in Rural Bangladesh
  Rayhan Rashed et al. | ACM COMPASS 2025
  📄 PDF | Slides

• Dakter Bari: Healthcare Services for Extremely Impoverished People
  Rayhan Rashed* et al. | ACM CSCW 2021
  📄 PDF | Presentation

• Tensions between Uber Owners and Drivers in Bangladesh
  SM Taiabul Haque, Rayhan Rashed et al. | ACM CSCW 2021
  📄 PDF

See the complete list on my Google Scholar.

Background

Education
– BS & MS in CS, Bangladesh University of Engineering & Technology (BUET)
– 🏆🏆 Prime Minister’s Award for Best MS Research in ICT, Bangladesh (2024)

Industry
Research Engineer, National Center for Artificial Intelligence (NCAI) — Worked in developing KALEEM, LLM-powered personalized education platform with virtual tutors.

Prior Research During my undergrad, I extensively studied (CSCW 2021, COMPASS 2021) and built (CSCW 2021) stakeholder-within-loop systems for social-good.

Beyond Research I’m a husband 👨‍❤️‍👩 and Chelsea fan !

Feel free to reach out for collaboration, discussion, or just a chat!