Saikat's Projects
This project was done as a part of Amazon Business Research Analyst Hiring Challenge. Here I was provided an dataset for the candidates for a XyZ company. I had to make prediction about the FitmentPercent. Also, the prediction had to be free from any bias. On the predicted Fitmentpercent, I had to predict if there was any bias and if there is bias then what kind of bias was present. By solving this project, I was qualified for the round 2.
This is part of hiring process for Amazon Business Analyst. The data provided here has two target variable - BiasInfluentialFactor and FitmentPercent. The prediction should be bias free as per requirement.
An implementation of the TKDE paper "Bidding Machine: Learning to Bid for Directly Optimizing Profits in Display Advertising"
We'll try to predict the buyers time spend on an e-commerce site
HackerEarth Machine Learning Challenge: Carnival Wars! Our task is to predict the selling price of the products based on the provided features.
Different cloud ops files for work
Being a part of the data engineering team, are expected to “Develop input features” for the efficient marketing model given the Visitor log data and User Data. As a Data Engineer Creating ETL Pipeline is expected.
first deployment trial
testing to run a docker file
A curated List of Coding Questions Asked in FAANG Interviews
This is hiring hackathon. Retail price prediction challenge by machinehack.com. Challenge to come up with an algorithm to predict the price of retail items belonging to different categories.
This is multiclass classification problem. In this problem, we are given a dataset that contains grievances of various people living in a country. Our task is to predict the importance of the grievance with respect to various articles, constitutional declarations, enforcement, resources, and so on, to help the government prioritize which ones to deal with and when.
Predict house price hackathon by machinehack.com
Predict the promotion of an employee
Analytics Vidhya hackathon problem. This is a classic imbalanced classification problem where we have to predict credit card lead. Resample and threshold shifting strategy used to increase the accuracy
Python library to extract important features from a dataframe
Large Language Model Projects
LTFS has tasked us with building a model given the Top-up loan bucket of 128655 customers along with demographic and bureau data, to predict the right bucket/period for 14745 customers in the test data.
This repository contain different modules of nlp
Grid Search, Pipeline, Bayesian optimization, hyperopt and optuna has been coded
Testing different ML models on Titanic Dataset
Connect APIs, remarkably fast. Free for developers.
Topped 61 out of 2100+ participants in HackerEarth Machine Learning Challenge. Problem was to predict the power that is generated in kW/h provided in the dataset
Predict the churn score for a website based on the features provided in the dataset. Ranked 70/4000 in Hackerearth
Config files for my GitHub profile.
This datasets is related to red variants of the Portuguese "Vinho Verde" wine. The data has a field called "quality" which has a range between 1-10, we'll assume 1-5 score is for bad wine and 6-10 is for good wine and we'll try to to predict the same based on the data.