Figarri Keisha 🤖
Figarri Keisha

MSc AI for Sustainable Development Student

University College London

About Me

A results-driven data scientist with 6 years of experience currently pursuing an MSc in Artificial Intelligence for Sustainable Development at University College London. I specialise in developing impactful AI solutions for sustainable development, particularly in education. With a keen interest in enhancing large language models and advancing AI fairness, I aim to bridge technology and societal impact to drive meaningful change.

Interests
  • Factually Grounded Text Generation
  • Evaluation & Interpretability of LLMs
  • Low-Resource NLP
  • LLMs for Developmental Policy Impact
  • Ethical & Safe Deployment of Foundation Models
Education

  • MSc Artificial Intelligence for Sustainable Development

    University College London

  • BSc Physics

    Institut Teknologi Bandung

Featured Articles
An Automatic Document Validation System in the Platform Merdeka Mengajar by Using Word Embeddings and Supervised Classification Models
🌾 Empowering Local Farmers with Data - KaPaN
📈 Causal Inference - Can We Trust What Happens After?
Recent Publications
Figarri Keisha, Zekun Wu, Ze Wang, Adriano Koshiyama, Philip Treleaven (2025). Knowledge Collapse in LLMs: When Fluency Survives but Facts Fail under Recursive Synthetic Training. arXiv.
Cite DOI URL
Figarri Keisha, Prince Singh, Pallavi, Dion Fernandes, Aravindh Manivannan, Ilham Wicaksono, Faisal Ahmad (2025). All for law and law for all: Adaptive RAG Pipeline for Legal Research. arXiv.
Cite DOI URL
Figarri Keisha, Andika R. Hakim, Bagoes R. Widiarso, Raden Z. Hermawan, Aghnia M. Safira, Rizka Azmira, Putri W. Novianti (2025). An Automatic Document Validation System in the Platform Merdeka Mengajar by Using Word Embeddings and Supervised Classification Models. Data Science and Artificial Intelligence.
Cite DOI
Recent Projects
See all