Experiences
Overview of My Professional Journey.
Created an Excel add-in "GoToColumn" using TypeScript to let clients access and manipulate columns easily.
Utilizes the Office JavaScript API for interacting with strongly-typed Excel objects including worksheets.
Developed a system to parse and consolidate bank statement PDFs into Excel using Python and Regex.
Increased productivity by reducing manual lookups through replacing Power Query Editor with Python code.
Projects
Here's what I have built throughout my journey.
Crafted a robust real-time collaborative platform with integrated chat, coding, and sketching features.
Implemented auto-language detection and in-browser code execution using CodeMirror component.
Enabled WebSocket communication for instant messaging and real-time notifications on join/leave events and activities, ensuring security and integrity.
User-friendly application for API testing and interaction, offering streamlined workflows for newbie developers.
Implemented request history and repeat requests features, making API testing more intuitive.
Provided code snippet generation for Axios and Fetch, simplifying coding with ready-to-use templates.
Enhanced Excel user experience with advanced column navigation and management, built using Typescript.
Significantly reduced time spent accessing columns across spreadsheets by over 50% while boosting accuracy and productivity.
Implemented features include hide/unhide, sheet locking, auto-refresh, and column profiling via Excel JavaScript API.
Developed a Streamlit application for file upload, text extraction, and conversational querying using an AI chatbot.
Text extraction using PyPDF2 and python-docx, text chunking for OpenAI GPT-3.5 model input limitations
Designed a user-friendly interface for seamless interaction and improved user experience
Pinpointed key factors impacting student success, using data visualization libraries and analytics.
Analysed patterns linking academics, activities, internships, etc, for better career choices.
Achieved an impressive accuracy of 0.81 by implementing advanced ensembling techniques, contributing to enhanced predictive modelling capabilities for student success initiatives.
Team project for showcasing university-wide events and handling room booking.
Developed user and admin signup/login page and linked it with the backend to store credentials in encrypted form.
Developed infinite scroll sponsorship section for enhanced user experience.
Research And Publications
Overview of My Research Journey.
Semantic Textual Similarity with Supervised and Unsupervised Learning
Developed a hybrid model combining Supervised and Unsupervised Learning approaches for Semantic Textual Similarity (STS) tasks.
Utilized SVR, LightGBM, and XGBoost models, alongside feedforward neural networks, to achieve enhanced text similarity predictions.
Achieved industry-standard performance with a Pearson correlation of 0.84, Spearman correlation of 0.82, and MSE of 0.73 by leveraging novel ensemble techniques integrating SVR and Neural Networks.
Conducted experiments on benchmark datasets (SemEval 2012 Task 6), achieving robust cross-validation results across 5 folds for 24 candidates in 120 fits.
Pinpointed key factors impacting student success, using data visualization libraries and analytics.
Analysed patterns linking academics, activities, internships, etc, for better career choices.
Achieved an impressive accuracy of 0.81 by implementing advanced ensembling techniques, contributing to enhanced predictive modelling capabilities for student success initiatives.
Tech Stack
Languages, libraries and frameworks I know and use.