I'm Working here as Senior Data Scientist in R&D department of excelra.com. I'm using here NLP technology and extensively working on GenAI projects. Leveraging the latest LLM techniques to get the best of technological advancement in NLP techniques.

- Led DS projects handling core business areas like Candidate Categorization System, Resume Parser, Skill recommendation, contributing to hiring, scaling and mentoring team, directly working with CEO, CTO and product head.- Candidate Categorization System - Improved models to categorize candidates on the basis of key attributes. Used LLM to extract key patterns thereby enhancing the candidate-job matching.- Resume Parser - Improved resume parser model efficiency by ~200% using BERT and LLM which was core service to extract candidate information. Worked on generic resume parser to handle varied resume formats.- Skills Recommendation - Used Word2Vec to model skill recommendation systems.- Job Title Prediction - Candidate profile data used to predict the Job title using word2vec, deep learning helping candidate to categorize into correct group

Image segmentation - Logo detection, Identifying signature and Identifying handwritten words.PDF Digitization - Converting scanned pdf to digital pdf such that it is searchable, more compactOCR - Training OCR to extract text from images, fine tuned on required data and also supported 6 languages.

Contributed to OCI Data Science development efforts, anomaly detection, explored NLP and AI services.

Chatbot - Developed a chatbot from scratch which predicts diagnosis on the basis of a series of symptoms using a decision tree and has also done feature engineering and data analysis on patient data.Technologies Used: Data preprocessing, Decision tree, Python, NLP, Feature engineering, Flask, Neural Network. etc.Developed a chatbot for Lactation assessment of mother after birth of child using NLP, advanced neural networks.

Implement customer categorization to target customers clusterwise having the same behavior and pattern for E-commerce: Used K-means clustering at E-commerce scale to cluster customers based on multiple factors like RFM score and categorize them into appropriate groups for targeted discount and customer retention.Sale prediction model for Inventory Management: Used Time series analysis(ARIMA) to predict by-weekly sale quantity by category.Recommendation System: Implemented Market Basket analysis using Apriori algorithm at E-commerce scale to suggest items to customers at cart based on existing cart items.Sentiment Analysis using NLP, LSTM to predict the overall sentiment of the customer from the website and social media.Identify the handwritten digits using Deep Learning( Convolutional Neural Network, Keras)

Building real-estate management service, working on Elasticsearch implementation for indexing, PDF to HTMLparser using PDFSharp library, calendar reminder features etc.