OJAAS S HAMPIHOLI

Data Scientist, ML Engineer, NLP Engineer, Data Engineer

  • Master of Science in Data Science - Indiana University Bloomington (2019 - 2021)

  • Bachelor of Engineering in Electronics and Telecommunication - University of Mumbai (2015 - 2019)

WORK EXPERIENCE

TECHPOINT SOS

Data Scientist | Indianapolis, IN | June 2020 - July 2020

  • Worked as a Data Scientist on a multi-disciplinary team to identify and implement solution for forecasting the COVID outbreak and detecting the counties most affected during COVID 19 pandemic.
  • Designed LSTM based Time Series Forecasting model and presented interactive maps using Tableau.
  • Worked remotely 20hrs per week to create the product prototype and helped with the go-to-market strategy.
  • Coordinated between business and technical team, thereby enabling a smoother communication between both teams.
  • Prioritized weekly work schedules for both business and tech teams by setting up and regulating the asana dashboard.
  • Designed the Product documentation and slide deck to be presented as a pitch idea to business investors.

FEATURED PROJECTS

Stock Forecasting

Implementation of Multivariate Vector Autoregression (VAR) and Deep Learning based RNN (LSTM) Model to predict the various categories of stock prices.

The model fitting for VAR is done using the BIC criterion and the forecasting is done for 15 days.

The Dataset obtained from the Alpha Vantage API is split into training and validation sets, which are used to train the model and to see its performance on unseen data.

To know more, click for the github link

Google Stock Prediction

Implementation of LSTM based Recurrent Neural Networks to predict univariate time series. The dataset used here is Stock Prices of Google obtained via Kaggle. The model has 4 LSTM layers followed by Dense layers and about 46 Mn trainable parameters.

The model achieves a Mean Absolute Error of 0.028 and predicts the prices very well. The visualization for the actual and predicted values are also included with the code.

To read more about the project, kindly have a look at the github link

Intel Image Classification Challenge

This challenge has images belonging to 6 classes namely Mountain, Street, Glaciers, Buildings, Sea and Forest. Our goal here is to build a model that detects the class of the given image.

CNN model fitting is done to predict the class of image with accuracy of 71%.

Transfer learning using VGG-16 model based on Imagenet is applied here to get higher accuracy of 83%

To know more, click for the github link

DCGAN

Implementation of Generative Adversarial Network to generate images of flowers.

The model is trained on the tf_flowers datsaet available in tfds.

The model generates images of flowers of dimension 300*300*3

DOMAIN PROJECTS

Machine Learning and Artificial Intelligence

Search Problems, Markov Models, Neural Networks and SVM, Supervised & Unsupervised Learning implemented on domains across Signal Processing, Computer Vision, NLP & Financial Data.

Big Data Modelling and Analytics

Relational and Non-relational Databases, Hadoop HDFS, Apache Spark

Multivariate & Exploratory Analysis

Single and Multivariate data analysis, Parametric and Non-parametric model fitting and analysis.

IoT and Embedded Systems

Design of Embedded Systems based on Arduino controller boards and IoT devices based on Raspberry Pi processors.

CERTIFICATIONS

SKILLS AND TECHNOLOGIES

The following section highlights the skill sets that I have acquired and the technologies that I have used over the years across various projects.
These skillsets pan across various domains including but not restricting to Data Science, Computer Science, Information Technology, Electronics, Telecommunication Networks

Sequence Models

Text Generation, Neural Machine Translation, Time Series Analysis to Predict Stock Prices, AR, MA, ARIMA, RNN's (LSTM's and GRU's)

Machine Learning

Various Supervised & Unsupervised Learning implemented on domains across Signal Processing, Computer Vision, NLP & Financial Data.

Artificial Intelligence

Search Problems, Markov Models, Neural Networks and SVM Designs.

Big Data Analysis

Big Data Pipeline Analysis, Apache Spark, Hadoop HDFS, MySQL, Oracle SQL & MongoDB

Statistics, Hypothesis Testing and Multivariate Analysis

T Tests, Chi Squared Tests, ANOVA, Multivariate Hypothesis Testing, MANOVA.

Exploratory Analysis

Univariate, Bivariate, Trivariate and Multivariate Data Analysis, Parametric and Non-parametric model fitting.

Signal Processing

Time and Frequency Domain Analysis, Signal & Control System Design, Image and Video Processing & Speech Processing.

Embedded Systems & IoT

System Designs, Testing and Analysis based on Processors (Raspberry Pi) and Controllers (Arduino).

ABOUT MYSELF

When I am not working, I am usually busy with reading novels and writing poems. I write poems in English, Hindi and Marathi.
I enjoy discussing new movies and listening to music. I also play the keyboard.
I am a Black Belt holder in Shotokan Karate (Certified by WFSKO), I like to spend my time playing Table tennis, Badminton and I love to watch Lawn Tennis Tournaments.

GET IN TOUCH

Contact Details