I am currently a 2nd-year MSc. Data Science student at RWTH Aachen University, Germany. I am specializing in the areas of Applied Machine Learning and Statistical Foundations of ML and AI. The focus areas of Computer Vision and Natural Language Understanding have highly intrigued me so far during my studies. Owing to the versatile and powerful applications of ML in every aspect of life, I lean towards an interdisciplinary approach of Applied ML.
Feel free to drop a text on LinkedIn (or an email) ^_^
Some cool updates (Click to expand)
- [November 2023] I receive the Deutschlandstipendium 2023, and funded by Porsche A.G.
- [June 2022] I graduate from college (BSc. Statistics) with honours and first class distinction
- [May 2022] I receive an admit from RWTH Aachen University
- [July 2019] I was awarded the INSPIRE Scholarship by the Ministry of Education (India) for achieving top 1% rank in the national ISC Examination
Some fun projects hidden behind Datenschutz :') (Click to expand)
- [December 2023]
Variable-lag identification for multiple time-series analysis using improved dynamic time warping : Dr. Ximeng Cheng (Department of AI, Fraunhofer HHI, Berlin)
- Identified dynamic lag in autocorrelated time series
- Performed Causality analysis (Granger's, etc.) to obtain information about lag effect
- Developed an implementation Dynamic Time Warping with moving window and correlation analysis (along with Transfer Entropy)
- Improved forecasting performance with feature engineering
- [September 2021]
Hierarchical Clustering in Jet Reconstruction Algorithms : Prof. Dr. Debarghya Ghoshdastidar, Dr. Stefan Kluth (TU Munich with Max Planck Institute for Physics)
- Interpretability of hierarchical clustering algorithms in data obtained from hadronization
- Implemented 7 different clustering algorithms β Sequential Clustering, Average Linkage, Ward Linkage, DIANA (or Divisive-Analysis) with pairwise distance, DIANA with JADE/Durham distance, Recursive Spectral Maxcut, and Recursive Normalized Spectral Maxcut, to simulated and real data obtained from hadronization on the hadron, parton and data level
- Developed a modified distance metric to perform comparative analysis of the clustering results
- Generated algorithms that are Collinear Safe, developed a new metric to quantify performance
- Developed global clustering algorithms inspired from the k-means++ technique
- [March 2021]
Exploratory Data Analysis on Time Series Sales Data : Prof. Dr. A. Chandra, Prof. Dr. D. Bhattacharya (St. Xavier's College, Kolkata)
- Analyzed 4 years of sales data (2 million+ data points) of a leading jewellery brand in India
- Conducted time series analysis on the data along with developing relevant business insights
- Used R programming language for data-cleaning, data pruning and data-visualisation
- Performed Customer Segmentation using RFM Model and k-means Clustering