Anirban Banerjee

ML Engineer | AI Developer | Data Scientist

NIT Rourkela '27

BTech in Ceramic Engineering

Minor in Computer Science and Engineering

Anirban Banerjee Photo

About Me

Hi! I am Anirban Banerjee. Prefinal year BTech student at NIT Rourkela. I am a AI ML engineer and I like to develop things using machine learning and AI which will have a real world imapact. My work bridges deep learning, computer vision and natural Language processing, with focus to solve day to day problems. I have developed multiple projects which involves reccommender systems, time series forcasting, vision transformers and language models.

I am comfortable with pyTorch, Tensorflow and other ML frameworks. I am also eager to learn, explore and contribute to developing libraries with bright future prospects. With a strong foundation in both theoretical and applied machine learning, I would like to collaborate on research and projects involving CV, NLP or llm which can open new pathways for the development of AI.

Education

Ramakrishna Mission Vidyalaya, Narendrapur

National Institute of Technology, Rourkela

Achievements

Interests and Specializations

Machine Learning
Machine Learning Algorithms
DL
Deep Learning
NLP
Natural Language Processing
AI
Artificial Intelligence
DSA
Data Structures and Algorithms

Projects

VAKYA: An Indian Language Text Segmentation model for NLP tasks

Indian Languages are often diversified and resources are low. Often meaning and context has to be preserved for translation, text summarization and other NLP tasks. So this model makes this process easier by using heuristics approach, generating embeddings using transformers and graph clustering algorithms for better segmentation.

View on GitHub
Text Recognition from old-middle age spanish documents

Traditional OCR methods often fail to recognise text from old documents due to ink bleed, noise, and old Languages. This project includes noise removing by Opencv methods, text detection avoiding unnecessary items in the documents. Then Text Recognition using vision transformer models. I also included various methods to translate the obsolete letters, replacing them with the letters which are in use now, with the help of Spanish Dictionary. This model has a 24% Character Recognition Error (CER) and gives a F1-score[Charcter Level] of 84%.

View on GitHub

Research

Natural Language processsing, deep Learning

I would love to contribute on research projects that involves NLP, DeepLearning.So, if you loved my works feel free to contact me. I would like to contribute to research project as much as I can.

LLMs , Foundation Models

Recent developments in the field of AI has led to the vast development of foundation models and LLMs. I am learning this field and also exploring the user aspects of it. I would love to collaborate on research projects which will open new aspects in this field.

Resume

view my resume here

Contact

Email: anirbanbanerjee3103@gmail.com

GitHub LinkedIn