CV
My full curriculum vitae, including education, research experience, publications, and academic service.
Contact Information
| Name | Ayush Lodh |
| Professional Title | Master's Student in Computer Science & Engineering |
| ayushlodh26@gmail.com | |
| Location | New Delhi, Delhi |
Professional Summary
Master’s student at NIT Delhi researching computer vision, deep learning, and pattern recognition, with a focus on handwriting recognition and document understanding. Author of peer-reviewed papers at ICDAR 2026, ECML PKDD 2026, ACPR 2025, and ICDAR 2025; past research intern at CVPRU (ISI Kolkata) and the Computer Vision Center (UAB, Barcelona).
Experience
-
2025 - 2026 Barcelona, Spain (remote)
Research Intern
Computer Vision Center, Universitat Autònoma de Barcelona
- Worked on multimodal retrieval-augmented generation with graph neural networks over enterprise document images, advised by Dr. Sanket Biswas and Dr. Josep Lladós
- Developed From Chunks to Graphs, a training-free multimodal retrieval framework that replaces chunk-based selection with structure-aware subgraph retrieval over Doc2Graph representations, improving factual consistency on document VQA over text-RAG baselines — published at ICDAR 2026 (first author)
-
2024 - 2024 Kolkata, India
Junior AI Engineer
Habitat Lens Private Limited
- Worked on handwritten document images and annotation pipelines for AI solutions
- Co-developed NoTeS-Bank, a benchmark evaluating vision-language models on handwritten academic and scientific notes, working with Alloy Das — published at ECML PKDD 2026
-
2024 - 2024 Kolkata, India
Research Intern
Computer Vision & Pattern Recognition Unit (CVPRU), Indian Statistical Institute
- Worked on online and offline handwritten text recognition from live camera feeds, advised by Prof. Umapada Pal
- Proposed an end-to-end transformer for writer-independent handwriting recognition fusing offline image patches and online stroke tokens via learnable latent queries — published at ACPR 2025 (first author)
Education
-
2025 - present New Delhi, India
M.Tech
National Institute of Technology Delhi
Computer Science & Engineering (Analytics)
- CGPA: 7.5/10
- Advisors: Prof. Dr. Nisha Singh Chauhan (NIT Delhi) and Dr. Sanket Biswas (CVC, UAB, co-supervisor)
- Project: From Chunks to Graphs — published at ICDAR 2026 (first author)
- Coursework: Data Mining and Warehousing, Deep Learning, Advanced Databases
-
2020 - 2024 Burdwan, India
B.E.
University Institute of Technology, The University of Burdwan
Information Technology
- CGPA: 8.79/10
- Final-year project: 3D Playfair Cipher with Dual Key (advisor: Ms. Nabanita Das)
- Coursework: Image Processing, Artificial Intelligence, Machine Learning
Skills
Computer Vision & Document AI (Advanced): Handwriting Recognition, Document Understanding, Multimodal RAG, Visual Question Answering
Deep Learning (Advanced): PyTorch, TensorFlow, Transformers, Vision-Language Models, Graph Neural Networks
Programming (Advanced): Python, C/C++, Git, Linux
Documentation: LaTeX, MS Office
Languages
Bengali : Native
English : Fluent
Hindi : Fluent
Interests
Computer Vision: Document Image Analysis, Handwriting Recognition
Pattern Recognition: Writer-Independent Online and Offline Handwritten Text Recognition
Deep Learning & Machine Learning: Transformers, Vision-Language Models, Graph Neural Networks
Document Understanding: Multimodal Retrieval-Augmented Generation (RAG), Document Visual Question Answering
Open Source: PyTorch, TensorFlow, HATChar-Classifier