DEEPTI VERMA

DEEPTI VERMA

0 connections
Connect
About
Data Scientist specializing in machine learning, time-series forecasting, and NLP. Built predictive models using XGBoost, LSTM, ARIMA, ARIMAX, and Random Forest on large-scale datasets, and developed RAG-based chatbot systems using LangChain, FAISS, and LLaMA-3. Skilled in feature engineering, statistical modeling, and model evaluation.
Experience
Data Science Intern (Time Series Analysis)

Energy Analytics Lab, IIT Kanpur

May 2025 - July 2025

- Evaluated 250K+ DAM electricity price records to identify temporal patterns and market volatility. - Performed feature engineering (lags, rolling statistics, external variables) to improve model performance. - Assessed anomaly detection techniques to identify and handle outliers in electricity price data. - Improved forecasting models (LSTM, ARIMA, SARIMA, XGBoost), with XGBoost achieving the lowest RMSE.

Education
M.Tech

Indian Institute of Technology Kanpur

August 2024 - Present

Field: Management Sciences

B.Tech

Rajiv Gandhi Institute of Petroleum Technology

July 2016 - May 2020

Field: Petroleum Engineering

Skills
Advanced
ARIMA ARIMAX LSTM NLP Python Random Forest Scikit-learn SQL XGBoost
Intermediate
Databricks Power BI PySpark PyTorch TensorFlow
Projects
AI-Powered Educational Chatbot (RAG)

January 2025

Built a RAG chatbot using LangChain, FAISS, and LLaMA-3 for transcript-based question answering.

Food Ordering NLP Chatbot

January 2025

Built an NLP chatbot using Dialogflow for intent classification and entity extraction.

MNIST Image Classification using Semi-supervised GAN

January 2025

Built CNN for MNIST classification achieving 98% accuracy on full data; dropped to 77.2% with limited data.

Customer Churn Statistical Modeling

January 2024

Utilized hypothesis testing and Logit/Probit regression to find key drivers of customer churn.

Contact Information
Quick Stats

0

Connections

0

Posts

1

Experiences