Kirandevraj R

Kirandevraj R

Data Scientist at Verisk Analytics

M.S. by Research, IIIT Hyderabad

LinkedIn GitHub Google Scholar

I work at the intersection of computer vision, speech processing, and NLP. My research focuses on fine-grained visual recognition, keyword spotting, and sign language understanding under limited supervision. At Verisk Analytics, I build production ML pipelines for OCR, image classification, and LLM-based text analytics for risk assessment.

Publications

  1. Dense Frame Annotations for Low-Resource ISL Fingerspelling Recognition.
    Kirandevraj R, Vinod K Kurmi, Vinay P Namboodiri, C.V. Jawahar.
    Under review, 2026.
  2. Continuous Fingerspelling Dataset for Indian Sign Language.
    Kirandevraj R, Vinod K Kurmi, Vinay P Namboodiri, C.V. Jawahar.
    AACL IJCNLP Workshop on Sign Language Processing, 2025.
    paper project
  3. Multilingual Query-by-Example KWS for Indian Languages using Transliteration.
    Kirandevraj R, Vinod K Kurmi, Vinay P Namboodiri, C.V. Jawahar.
    INTERSPEECH (Oral), 2025.
    paper code
  4. Generalized Keyword Spotting using ASR embeddings.
    Kirandevraj R, Vinod K Kurmi, Vinay P Namboodiri, C.V. Jawahar.
    INTERSPEECH (Oral), 2022.
    paper code

Education

International Institute of Information Technology, Hyderabad Aug. 2019 - May 2022
  • M.S. by Research in Computer Science and Engineering
  • Advisors: Prof. C.V. Jawahar, Prof. Vinay P Namboodiri, Prof. Vinod K Kurmi
Valliammai Engineering College, Anna University Aug. 2013 - Jun. 2017
  • B.E. in Computer Science and Engineering

Work Experience

Verisk Analytics Hyderabad, India | Jun. 2022 - Present
  • Deployed DeepDocDetection OCR pipeline to extract business information from documents.
  • Deployed BLIP-2 and CLIP models via MLflow for zero-shot image classification and fine-grained object detection for risk assessment.
  • Built end-to-end LLM-based pipelines using AWS Bedrock and RAG architectures for text analytics.
Wipro Bangalore, India | Oct. 2017 - Jan. 2019
  • Worked on SAP software to provide Enterprise Resource Planning service to Marathon Petroleum Corporation.

Teaching Assistant

Software Engineering for Data Science IIIT-H | Jun. 2021 - Sep. 2021
  • Conducted TA sessions on Machine Learning (Supervised, Unsupervised, NLP) and Big Data technologies (Hadoop, Spark, Hive, Kafka).

Academic Service

Reviewer for ICASSP 2025, 2026