Reddit Data Analytics

Reddit Post-Comments Relevance analysis

• Skilled at pinpointing the pertinent comments using pre‐trained LLM’s for specific Reddit posts or questions, along with the ability to identify relevance.
• Integrated statistical Topic modeling ( Latent Dirichlet allocation) with the LLM models like Bert, MiniLM, GPT, BertQuora, also known as Hybrid Models, to boost the performance of the contextual semantic analysis by 30%.

Home Credit Default Risk (HCDR)

HCDR Project

• Utilized Kaggle dataset , performed data collection, wrangling, feature engineering, exploratory data analysis(EDA), Interactive Visualisation Dashboard and Predictive analysis.
• Technology used - Python, EDA, Feature Engineering, ML Model(XGBoost, SVM), Neural Networks
• Implemented data engineering techniques to clean and prepare raw data using ER chart for analysis, ensuring accuracy and reliability in predictive modeling for credit risk assessment

Data Science Report

Part of IBM Data Science Professional

SpaceX Data Science Capstone Project

• Given data set, performed data collection, wrangling, feature engineering, exploratory data analysis(EDA), Interactive Visualisation Dashboard and Predictive analysis.
• Trained and evaluated ML models, the outcome is, the best model - Decision Tree with accuracy - 87%
• Able to understand the business needs and perform technical analysis or evaluations using Data Science practices.

MNIST Hand written digit recognition

First Neural Network project

Handwritten digit recognition Project

• MNIST dataset
• Handwritten digit recognition using simple neural network, with MNIST dataset
• Able to understand the Neural Networks and its structure using tensors.

Live Covid/Visualisation Dashboard

Built during Covid - So that people can visualise the areas affected and plan accordingly

• Performs live web scraping for every access operation, built on the flask.
• Built an Interactive Geospatial visualization and choropleth map on folium.
• Dash framework is used in the interface (for interactive dashboard).
• Researched and implemented statistical measures and displayed them in the website.
• As in 2022, Tried hosting in Aws instead of Heroku, for better performance.
• Currently learning Docker and Kubernetes for better performance and low latency. ----- 50% (Docker done)

Portfolio

Finally! This is my first Web Portfolio you are viewing 😍

Portfolio

• A portfolio built using HTML, CSS, Bootstrap, and Javascript.
• Domain name forwarded to GitHub, using Github as my hosting platform.
• Continously updating, currently looking forward to build using React
• In 2022, Deployed blockchain website with the help of IPFS hash.
• Blockchain Domain - Mr-King.nft
• Above website is accessible using Unstoppable.
• Technology shift from web2 to web3

Final Year Project

Team of 2 - under supervision of Professor Dr.Dithipriya sinha,

Title - Health Care System based on Blockchain Technology

• EHR (Electronic Health records) are stored in a Blockchain having many set of constraints to maintain security and accessible only to the un-restricted users (particular patient and specific Doctor who evaluate and assess their medical condition in prescription format) prescriptions are stored in separate Blockchain network. Which is easily accessible to pharmacist.
• •Team of 2, I have done the research and initiated the Geth(Go Ethereum) Blockchain in the local host and programmed basic smart contracts with respect to idea behind.
• As per prior experience from training and internships, programmed IoT device and built machine learning model for predictions.
• Designed architecture and implemented an end-to-end application with security constraints.

Final Year Project

Team of 3 - under supervision of Professor Dr.Dithipriya sinha,

Masters scholar - Raj Vikram

Title - Blockchain based FireForest prediction (Dapp)

• I'm a Team lead
• Things I have done - (mostly backend)
• Developed a prediction process using datasets - Python,scikit-learn,metrics accuracy
• Machine learning algorithm used linear regression

---

Blockchain Part -

• Researched and developed the blockchain private network in local host using Go ethereum(Geth)
• Performed transactions, mining process and p2p( node <--> node / in Simple terms interprocess communication) communication was established
• Deployed Smart Contract, during deployment explored various communication processes like rpc, https, Javascript- wget, localhost
• Smart Contract was built in solidity and developed in Remix IDE
• Finally, Built and deployed an API using Web3 for communication between frontend <--> backend (entire backend was based on nodejs)

Mini Project

Team of 3 - under supervision of Professor Dr.Dithipriya sinha,

Masters scholar - Raj Vikram

Title - IoT Sensor based on Blockchain Technology ( Fire detection Alarm)

• It is a dapp basically alerts all the neighbors and fire station. It has a capability to predict the occurrence of fire accidents in a factory or an organisation./li>