Still Struggling with Tech Hiring? Discover Olibr's Solution Now!
Login / RegisterDeveloperClient
Login/Register

Hargurjeet Ganger

Data Scientist

  • Bengaluru, KA
  • 11 years
Notice Period: 90 Days

    Candidate's About

    I am a data science professional with experience in statistics, data exploration, machine learning, and Azure. I actively engage in open-source platforms and have a blog and YouTube channel to educate others on ML and data science concepts.

    Work & Experience

    D

    Data Scientist

    Sep 16 - PRESENT

    Technology

    • Icon
      SQL
    • Icon
      Azure
    • Icon
      Python
    • Icon
      Machine Learning
    • Icon
      Problem Solving
    • Icon
      Statistics
    • Icon
      PyTorch
    • Icon
      Power BI
    • Strong experience in Python for data wrangling/analysis using Pandas, Numpy, Matplotlib, seaborn, etc.
    • Built Power BI dashboard to forecast Materials on-time delivery in 5 geographies within 1 year, resulting in a 10% reduction in budget allocations.
    • Developed ML models (XGBoost, Random forests, Decision trees) for predictive maintenance of equipment at oil refineries, resu...
    Read More

    Zomato Restaurant Analysis

    Exploratory Data Analysis and Visualization

    • Conducted analysis on a dataset comprising over 9000 restaurants, spanning 15 countries.
    • Utilized Python, specifically pandas and NumPy, for data analysis and manipulation.
    • Employed Matplotlib and seaborn libraries to create visualizations and gain insights from the data.
    • Identified top locations for 'Breakfast and Coffee,' 'Fast Food' restaurants, and 'Ice Cream' parlors in different localities based on the analysis results.


    • Python
    View Details

    Nifty 50 Price analysis

    Time Series forecasting

    • Analyzed 20 years of price history and trading volumes data for Nifty 50 stocks in the Indian stock market.
    • Implemented moving averages techniques to identify trends and patterns in the data.
    • Utilized augmented Dickey Fuller Test and stationarity conversion techniques to assess the stationarity of the data.
    • Built models using Arima and Prophet algorithms to forecast future stock prices.
    • Evaluated the performance of the models by considering RMSE (Root Mean Squared Error) and MAE (Mean Absolute Error) metrics.


    • Propeht
    • Arima
    View Details

    Twitter Disaster Tweets

    Data Scientist

    • Processed textual data to determine the authenticity of tweets containing information about disasters.
    • Built a deep neural network using TensorFlow to perform the classification task.
    • Utilized the NLTK library for implementing techniques such as stop word removal, tokenization, pad sequence, and word embedding.
    • Applied stop word removal to eliminate common words with little significance, tokenization to convert sentences into individual words, pad sequence to ensure uniform length of input data, and word embedding to represent words as dense vectors.





    • TensorFlow
    View Details

    Fruit 360 Classifier

    Deep Learning Projects

    • Constructed a CNN (Convolutional Neural Network) using PyTorch to predict the class of fruits from a dataset of 65,000 images, encompassing 131 different classes.
    • Implemented various techniques to enhance model performance, including data augmentation, batch normalization, learning rate scheduling, weight decay, gradient clipping, adam optimizer, layer dropouts, and minmax pooling.
    • Trained and evaluated the model on a GPU using PyTorch's CUDA library for accelerated computations.
    • Achieved an impressive model accuracy of 98% in less than 5 minutes, showcasing the effectiveness of the employed techniques and the computational power of the GPU.


    • PyTorch
    View Details
    I

    IT Analyst

    Feb 10 - Aug 16

    Technology

    • Icon
      Python
    • Icon
      SQL
    • Icon
      Azure
    • Icon
      Machine Learning
    • Icon
      Power BI
    • Icon
      Problem Solving
    • Icon
      Statistics
    • Icon
      PyTorch
    • Performed System & Integration Testing to validate client PoS and payment system.
    • Supported and managed User Acceptance Testing single-handedly working with multiple Business stakeholders.
    • Spent one year in the UK guiding offshore teams with the implementation of new PoS software.
    • Awarded "Star of the Month" twice.
    • Promoted from Systems Engineer to IT An...
    Read More

    Education

    N

    B.E in Electronics and Communication

    New Horizon College Of Engineering
    2006 - 2010

    Certificates

    M

    DP100 - Azure Data Scientist Associate

    Microsoft
    M

    DP900 - Azure Data Fundamentals

    Microsoft
    L

    Deep Learning Applications with Keras

    Linkedin
    C

    NLP using Tensorflow

    Coursera
    M

    AI102 - Azure AI Engineer

    Microsoft
    J

    Data Analysis with Python: Zero to Pandas

    Jovian.ai
    K

    Intermediate in Machine Learning

    Kaggle
    J

    Deep Learning with PyTorch

    Jovian.ai

    Achievements

    • Conducted a time series analysis and forecasting of the Nifty 50 stock prices and trading volumes, using historical data spanning 20 years. Implemented moving averages techniques, augmented Dickey Fuller test, stationarity conversion techniques, ARIMA, and Prophet models and evaluated their performances using RMSE and MEA errors.
    • Processed and classified large volume of twitter data for disaster-related tweets as fake or genuine using deep neural network built in TensorFlow and natural language processing techniques such as stop word removal, tokenization, pad sequence and word embedding using NLTK library.
    • Constructed PyTorch CNN for 65k fruit images, 131 classes. Employed data augmentation, normalization, LR schedule, Adam, dropout, Minmax pooling, reaching 98% accuracy in <5 mins.