Experience

๐Ÿฅ DICOM Automation Service (Under Prof. Dr. Deepak Agrawal, AIIMS Delhi)

Research Collaborator Jul 2025 โ€“ Sep 2025 ยท New Delhi, India

  • Architected and deployed an end-to-end video-to-DICOM pipeline using Python, PostgreSQL, and Dockerized microservices integrated with Orthanc PACS.
  • Achieved 95% workflow automation and near-real-time processing (approx 30s per video), significantly improving medical-image processing efficiency and clinical readiness.
  • Completed the research project which was recognized as a significant step toward improving medical-image processing within AIIMSโ€™ digital-health research environment.

๐Ÿ’ป IBM SkillsBuild | Agentic AI: From Learner to Builder

AI Agent Architect Training Program Jul 2025 โ€“ Aug 2025 ยท Remote

  • Completed a 4-week intensive training program on Agentic AI and AI Agent Architecture with the CSRBOX Foundation.
  • Gained hands-on experience in building and implementing production-ready AI agents using the Gemini API.

๐Ÿ”ฌ GGSIPU โ€“ University School of Automation and Robotics

Undergraduate Research Associate
Jul 2024 โ€“ Jun 2025 ยท New Delhi, India
Mentors: Dr. Sanjay Kumar Singh (GGSIPU), Dr. Pranshu Saxena (Bennett University)

  • Conducted deep learning research in medical and agricultural imaging.
  • Published 2 IEEE papers focused on transfer learning and Bayesian optimization.
  • Achieved 97.23% accuracy in tomato leaf disease detection using optimized Xception.
  • Enhanced breast tumor classification to 98.46% using fine-tuned CNNs and Population-Based Training.
  • Currently leading a journal-level study involving MC Dropout and temperature scaling for calibrated predictions.

๐Ÿ“Š M.K. Associates (Remote)

Financial Data Analyst Intern
Jan 2024 โ€“ Jun 2024

  • Performed EDA on 10K+ financial transactions across 700+ client accounts.
  • Identified and flagged 150+ high-risk anomalies, improving audit workflows.
  • Built ML models for financial risk classification (93% accuracy).
  • Automated reporting in Excel, reducing manual effort by 40%.
  • Supported regulatory compliance by aligning outputs with business policies.