Sai Panikar

Integrated RNS and INSID3 into a unified training-free segmentation pipeline, combining CLIP-and-FAISS-based retrieval with DINOv3 dense features for spatial localization.

Research Intern - Indian Institute of Science Education and Research Thiruvananthapuram - Thiruvananthapuram

(2026-05 - 2026-06)

Integrated RNS and INSID3 into a unified training-free segmentation pipeline, combining CLIP-and-FAISS-based retrieval with DINOv3 dense features for spatial localization.

• Adapted and extended existing implementations of both methods to work together within a single pipeline, rather than as standalone components
• Evaluated the integrated pipeline on in-context and multi-class segmentation tasks across ADE20K and PASCAL VOC benchmarks
• Gained hands-on experience working with published research codebases, dense visual feature extraction (DINOv3), and retrieval-augmented pipelines (CLIP + FAISS) for spatial reasoning tasks

Co-ordinator - Indian Institute of Technology, Roorkee - Roorkee

(2025-10 - 2025-10)

Served as a Coordinator for Thomso, overseeing planning and execution of assigned events, coordinating with team members and participants, and ensuring smooth on-ground operations during the 3-day fest.

M.Sc.

• Mathematics - Indian Institute of Technology, Roorkee (2024 - 2026)

Graduate (UG) - Dr Dy Patil ACS College (SPPU) (2021 - 2025)

Intermediate (Class XII) - SNBP Jr College (Maharashtra State Board) (2020 - 2022)

Matriculate (Class X) - Alphonsa High School (Maharashtra State Board) (2018 - 2020)