Tanish Jain
Machine Learning Engineer | Computer Vision & Deep Learning Research
Computer Vision and Deep Learning researcher specializing in object detection, medical image analysis, and scene understanding.
Research & Background
Pursuing M.Tech in Signal and Image Processing at NIT Rourkela (GPA 9.7/10), building on a B.Tech in Electronics and Communications Engineering from Assam University (GPA 8.2/10).
Completed a Software Developer Internship at Samsung R&D Institute India, Bangalore (Jan 2026 – Jul 2026), where I developed a zero-shot CLIP-based out-of-distribution (OOD) detection pipeline for live screenshot analysis.
PG Researcher at the PRISM Research Group, NIT Rourkela under Dr. Sobhan Kanti Dhara, conducting cutting-edge research in aerial image object detection and multimodal computer vision.
Co-authored publications across leading IEEE and Springer venues (including IEEE SPACE, IEEE GRSL, NCC, CVIP) and won the Best Paper Award at IEEE SPACE 2026.
Where I've Worked
Software Developer Intern
Samsung R&D Institute India
- Built a zero-shot CLIP-based out-of-distribution (OOD) detection pipeline for live screenshot images, applying image processing and vision-language techniques to guide Samsung Screen AI data curation.
- Extracted OCR from screenshot images, modeled text semantics with language models, and fused OCR-language and vision-language scores into a unified OOD score.
- Improved computational efficiency through statistical patch/region selection, feature sparsification, and local-global image feature aggregation for real-time analysis.
PG Researcher (under Dr. Sobhan Kanti Dhara)
PRISM Research Group, NIT Rourkela
- Primarily conducted research on aerial image object detection in optical satellite imagery under complex conditions.
- Worked as a support researcher on tiny object detection and thermal-aerial detection tasks in multimodal settings.
- Authored and architected GeoStatNet (IEEE SPACE 2026) and GFCR-Net (IEEE GRSL, under review), and co-authored SAPNet, TSRNet, and "From Pixel to People".
Freelance Deep Learning Engineer
Upwork
- Built ML pipelines for gigapixel whole-slide pathology images (SVS, ~80K–100K x 80K–100K pixels) using Multiple Instance Learning with ResNet features and attention pooling for tumor classification.
- Developed a cell-level mutation detection pipeline using segmentation masks to identify mutated cells in pathology imagery.
Data Science Intern
Navodita Infotech
- Developed a recommender system engine for an e-commerce platform using TensorFlow Recommender System.
- Implemented machine learning algorithms to analyze user preferences, personalize user experience, and improve recommendation accuracy.
Computer Vision Intern
MNIT Jaipur
- Built a proof-of-concept web application for live PCB defect detection using a deployed YOLOv7 image processing pipeline, image uploads, and real-time model inference.
- Added an interactive canvas for labeling missed defects, enabling faster feedback collection and dataset updates.
Featured Research & Projects
Research implementations, computer vision models, and deep learning systems. Click any project card to view full details.
Research Papers & Awards
Awards & Honors
Skills & Technologies
Certifications & Courses
Verified courses and specializations completed in Deep Learning and Data Science.
Let's Connect
Whether it's a research collaboration, full-time opportunity, or technical discussion — my inbox is always open.