M.TECH · SIGNAL & IMAGE PROCESSING · NIT ROURKELA
Tanish Jain
I build computer-vision models and practical AI software — from oriented ship detection and multimodal data-quality pipelines to local LLM, retrieval and coding-agent tools.
Experience
Industry, academic research and engineering work, 2023–2026.
About
A research-led engineer working across visual perception and applied AI software.
I am a signal-and-image-processing researcher who enjoys taking models beyond experiments — building, evaluating and engineering AI systems that solve concrete problems.
I am pursuing an M.Tech in Signal and Image Processing at NIT Rourkela, after completing my B.Tech in Electronics and Communication Engineering from Assam University. My work sits at the intersection of computer vision, deep learning and practical AI engineering.
During my internship at Samsung R&D Institute India, I worked on an MLLM-powered synthetic-data quality pipeline. I combined OCR signals, image features and vision-language embeddings, then used statistical out-of-distribution scoring to identify low-quality generated samples.
My research focuses on robust object detection in difficult remote-sensing imagery. I have developed GeoStatNet and GFCR-Net for small and oriented ship detection, and worked on 3D liver and tumour segmentation. This work has led to publications and research submissions in computer vision and remote sensing.
Outside research, I build usable AI tools: Priorscope for evidence-checked research discovery, HearthForge for reviewable local coding agents, and EdgePilot for local model inference and benchmarking. I care about measurable evaluation, reliable outputs and systems that remain useful beyond a demo.
- Current degree
- M.Tech · NIT Rourkela
- Undergraduate
- B.Tech · Assam University
- Industry
- Samsung R&D Institute India
- Research
- Ship detection · Medical imaging
- Engineering
- Multimodal AI · RAG · Local inference
- Based in
- India
Selected work
Selected work across vision, multimodal AI, inference engineering and LLM-powered applications.
Research
Research in robust visual perception, medical imaging and efficient deep learning.
Recognition
Toolkit
What I reach for, in the order I reach for it.
Certified
Let's
talk.
Research collaboration, a full-time role, or a question about a paper — the inbox is open.