Researcher

Kathan Pandya

Kathan Pandya is a bioinformatics researcher and a student assistant on the curATime project in the Data Science and its Applications group at DFKI, where her work spans systems biology, computational drug design and automated hypothesis generation. She studied for a BSc in Biological Sciences (Honours) at Krea University from 2020 to 2023, completed a computational drug design internship in 2022, and coauthored a 2023 Frontiers in Human Neuroscience paper on mild traumatic brain injury comorbidities. From 2024 to 2026, she studied for an MSc in Bioinformatics at Saarland University and coauthored BioDisco, a multi-agent biomedical hypothesis-generation framework that appeared on arXiv in August 2025 and in the IJCAI 2026 proceedings in August 2026. Her 2026 research also included coauthoring the March 2026 final reports for the biosignATure and curAIscid projects under curATime, as well as the July 2026 CleanSurvival preprint on reinforcement learning for survival-analysis preprocessing.

Now at
DFKI

Timeline

  1. 2020–2023

    BSc at Krea University

    She studies for a BSc in Biological Sciences (Honours) at Krea University from 2020 to 2023. During her undergraduate studies in 2022, she completes a summer internship focused on drug design using computational methods.

  2. June 2023

    A brain injury comorbidities paper

    In June 2023, she is one of five authors of a Frontiers in Human Neuroscience paper on comorbidities after mild traumatic brain injury. The study analyzes patient data from the TBI Model Systems national database.

  3. 2024–2026

    Bioinformatics master's at Saarland

    From 2024 to 2026, she studies for an MSc in Bioinformatics at Saarland University. Earlier experience includes an internship at a company that provides software solutions for biochemical problems and a research-assistant role in drug design and artificial intelligence.

  4. August 2025

    BioDisco on arXiv

    In August 2025, the preprint of BioDisco appears on arXiv with Pandya as third author, and a revised version follows in November 2025. The multi-agent framework combines language-model reasoning with biomedical knowledge graphs and automated literature retrieval to generate research hypotheses.

  5. March 2026

    Two curATime project reports

    In March 2026, she coauthors the final reports for two curATime projects with David Selby, Sebastian Vollmer and colleagues: biosignATure, on multi-omics biomarker signatures for atherothrombosis, and curAIscid, on AI science and development.

  6. July 2026

    CleanSurvival for survival analysis

    In July 2026, she coauthors the sixth version of the CleanSurvival preprint, presenting a reinforcement-learning framework that optimizes data preprocessing pipelines for survival analysis across continuous and categorical variables.

  7. August 2026

    BioDisco at IJCAI 2026

    BioDisco appears in the proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence in 2026, in the AI and Health track. The conference takes place in Bremen in August 2026.

  8. As of October 2026

    Student assistant at DFKI

    As of October 2026, she is a research assistant and student assistant at DFKI as part of the curATime project, in the Data Science and its Applications group. Her interests include systems biology and drug design and development.

Career

  1. NowDFKIStudent assistant

Education

  1. 2026Saarland UniversityMSc in Bioinformatics
  2. 2023Krea UniversityBSc in Biological Sciences (Honours)

Sources and credits

Photo: We have no published photo of Kathan Pandya yet. Send us a free-licensed photo.

  1. Kathan Pandya — Data Science and its Applications dsa.dfki.de. Checked 22 October 2026.
  2. Interwovenness: The fabric of exploration - Krea krea.edu.in. Checked 22 October 2026.
  3. Prevalence of comorbidities post mild traumatic brain injuries: a traumatic brain injury model systems study frontiersin.org. Checked 22 October 2026.
  4. BioDisco: Multi-agent hypothesis generation with dual-mode evidence, iterative feedback and temporal evaluation arxiv.org. Checked 22 October 2026.
  5. [2508.01285v1] BioDisco: Multi-agent hypothesis generation with dual-mode evidence, iterative feedback and temporal evaluation arxiv.org. Checked 22 October 2026.
  6. biosignATure - curATime: Systemorientierte, Multi-Omics-Identifikation von Biomarkersignaturen für die Detektion, Quantifizierung und Behandlung von Atherothrombose oa.tib.eu. Checked 22 October 2026.
  7. curAIscid - curATime AI science & development oa.tib.eu. Checked 22 October 2026.
  8. [2502.03946v6] CleanSurvival: Automated data preprocessing for time-to-event models using reinforcement learning arxiv.org. Checked 22 October 2026.
  9. BioDisco: Multi-Agent Hypothesis Generation with Dual-Mode Evidence, Iterative Feedback and Temporal Evaluation | IJCAI ijcai.org. Checked 22 October 2026.
  10. Kathan Pandya, B.Sc. www-live.dfki.de. Checked 22 October 2026.