Kunal Dutta
@user.2599194
Data Science Intern at Zigram India
Bengaluru, India
Kunal Dutta is a Masters in Data Science graduate from Christ University Bengaluru with a strong academic background in machine learning, NLP, and graph neural networks. He has research experience from the University of New Hampshire and Jadavpur University, and has worked as a Data Science Intern at Zigram India on PDF chatbots using Langchain and RAG-based architectures. His projects include recommendation systems using graph neural networks and COVID chest X-ray classification.
Experience
Data Science Intern
Zigram India Pvt Limited
Creation of PDF Chatbots: Worked on the creation of PDF Chatbots using Langchain Framework with the utilization of RAG Based Architecture. Used Unstructured API for efficient text structuring and embedding generation and Claude AI LLM to form a PDF Question Answering System. Concept Extraction: Worked on querying various Open Source Knowledge Graphs for searching linked terms to a given concept which was utilized in better search procedures relating to the term.
NLP Research Intern
University of New Hampshire
Entity Aspect Linking: Worked on design of an efficient algorithm using link prediction to correctly link the possible aspects of a content to the entities mentioned in a given context with the utilization of various Graph Neural Networks Architectures with Fine Tuned BERT Text Embeddings as Features.
Research Intern
Jadavpur University
Landslide Mortality Modelling through Geographically Weighted Regression(GWR): Worked on Spatial modelling of the effects of meteorological, topological and soil related factors on Landslide Mortality. Comparison was done on Normal GWR, Poisson GWR and Multiscale GWR. Modelling of PM2.5 through Geographically Weighted Regression(GWR): Worked on Spatio-Seasonal modelling of the effects of meteorological, socio-demographic and pollutant factors on PM2.5 levels on a seasonal basis.
Research Intern
University of New Hampshire
Knowledge Graph Reconstruction Analysis: Worked on the creation of Knowledge Graph Embeddings from benchmark Knowledge Graph datasets through various embedding methods such as HOPE, SDNE, LINE, Node2Vec followed by Graph Reconstruction to analyse the performance of the embedding techniques. Knowledge Graph Merging with Autoencoders: Worked on merging of Knowledge Graphs using Autoencoders to generate subgraph embeddings and intermediate meta embeddings followed by Knowledge Graph reconstruction for analysis of Meta embedding generation techniques.
Education
Christ University
Masters
Data Science
St. Xavier’s College(Autonomous)
Bachelors
Computer Science
Delhi Public School Ruby Park
CBSE AISSCE
Licenses & Certifications
Machine Learning Specialization
Coursera
Machine Learning A-Z
Udemy
Deep Learning Specialization
Coursera
Apprenticeship for Advanced Machine Learning and Deep Learning
Indian Statistical Institute
Statistical Inference with Python
Infosys Springboard