Headline

Computer Vision and HCI Researcher @ IISc | Multimodal AI & Adversarial Robustness | Autonomous Systems | Published at CVPR & ICCV Workshops

About

Applied Machine Learning and Computer Vision Engineer with 2+ years of research and development experience across premier institutes (IISc Bangalore, IIT Jodhpur, IISER Bhopal) and 3+ publications at prestigious venues including CVPR and ICCV workshops. I specialize in building end-to-end vision pipelines that operate reliably in messy, real-world environments. My technical focus centers on object detection (YOLO, DETR), Vision Transformers (ViT, ViViT), multimodal data synchronization, and Trustworthy AI, specifically making models resilient against Out-of-Distribution (OOD) data and physical adversarial edge cases. Passionate about solving real-world problems with AI, I’m eager to collaborate on impactful projects in the field of machine learning and computer vision.

Location

Bengaluru

AI stack

antigravity

claude

langchain

huggingface

ChatGPT

Work experience

Project Associate at Robert Bosch Centre for Cyber-Physical Systems @ IISc, Jul 2025 - Present · 1 yr 2 mos, Indian Institute of Science Bangalore, • Implemented a Time-to-Collision (TTC)-based Forward Collision Warning system within a 3-DOF driving simulator and developed a synchronized multimodal data pipeline integrating eye-tracking, heart-rate monitors, simulator telemetry, and cameras for large-scale multimodal data collection for driver behavior analysis. • Analyzed the cognitive load of 100 professional Indian heavy-vehicle drivers by conducting statistical analysis of ocular metrics (saccades, fixations, and pupil dilation) to evaluate cognitive workload fluctuations triggered by different ADAS alerts and complex driving maneuvers in heterogeneous traffic environments

Intern at Robert Bosch Centre for Cyber-Physical Systems @ IISc, Oct 2024 - Jun 2025 · 9 mos, IISc Bangalore India, Studied contrastive learning and its applications in computer vision, aiming to improve model performance through enhanced representations. Additionally, implement SimCLR on the CIFAR-10 dataset to evaluate its effectiveness for unsupervised representation learning and downstream classification tasks.

Product Developer and Researcher at Pune Institute of Computer Technology, Jun 2024 - Oct 2024 · 5 mos, Pune District, • Developed a computer vision model for recognizing signs in videos using a custom Indian Sign Language (ISL) dataset, employing MediaPipe landmarks and data augmentation techniques to reduce bias. • Achieved 95.6% accuracy with the Video Vision Transformer on the test set, following extensive hyperparameter tuning, and published findings at the 3rd International Conference for Advancement in Technology (ICONAT).

Master Thesis Student at Department of Computer Science and Engineering, IIT Jodhpur, Aug 2023 - May 2024 · 10 mos, Jodhpur, Rajasthan, India, • Created robust adversarial patches that maintain effectiveness across texture variations, targeting state-of-the-art object detectors like YOLO and Detection Transformer trained on diverse aerial datasets (DOTA, FGVC, COCO). • Extensively trained and evaluated various object detection models, implemented advanced adversarial attacks, analyzed their impact on detection performance, and explored novel attack strategies.

Summer Research Intern at iHub-Drishti, May 2023 - Aug 2023 · 4 mos, On-site

Physics Content Creator at Doubtnut, Dec 2019 - Jan 2020 · 2 mos, gurugram

Education

BS-MS, Data Science and Engineering, IISER Bhopal

Higher Secondary, Science, V. N. J. Inter College

Public links

github: https://github.com/ojojaswee