vrbabu9000 Goto Github PK

abusive_content_nlp

An advanced NLP project that checks for abuse in content. A strong contrast is drawn between baseline approaches and deep learning frameworks like BERT, LSTM.

advanced_ml_modeling

Well curated solution set for all ExcelR Data Science and Machine Learning problems which I had made during my time in the program

bank_termdeposit_predictor

Given a highly biased banking dataset of 45k entries and 17 variables wherein I was able to filter out good predictors for modelling and further balance the bias in the dataset using SMOTETomek method. The bias removal was approached in two different ways and the pros and cons of both methods are stated. The end result is a logistic regression model with 81% accuracy and striking a sharp balance between precision and recall with the former having a slight upper hand.

ey_better_working_world_frogs

forecasting_analysis_airline

Time Series analysis for airline dataset. Includes a general forecasting template which I made covering all forecasting techniques. Fast Pycaret Modelling for forecasting was also included

kaggle_titanic_statsmodeling

Taking on the first Kaggle Challenge: Titanic Disaster ML modeling, I had worked on advanced Feature engineering and Ensemble modeling techniques.

latencyslayerx

Reducing LLM API latency by implementing asynchronous multi-threading, allowing simultaneous execution and setting new operational standards.

scrapymongodb_numadic

This was a rush project I did while applying for an Internship at Numadic

structuredragnarok

A pun on "Ragnarok," the Norse apocalypse, implying a powerful, transformative approach to RAG with structured data and Watsonx.ai

vehipathoptimizer

This is an old project that I worked on back in Dec 2021 while trying to solve the Generic Vehicle Routing Fleet Optimization problem with Google OR

vrbabu9000