Circle Image

Jay Bhatt

Developer
“ Always code as if the guy who ends up maintaining your code will be a violent psychopath who knows where you live. ” - Rick Osborne

About Me

I am a graduate student at Stony Brook University studying Master's of Science in Computer Science.

I am best known as a Machine learning enthusiast. I like to dabble with new technology.

I recently interned at Google, in the Maps Team, at Mountain View.

Worked on creating a Pipeline to Detect and Transcribe Business Names in StreetView Images.

Areas of interest - Machine Learning, Data Science, Computer Vision, Cyber Security

Skills

Machine Learning Libraries

Scikit-Learn : 85

85

TensorFlow : 85

85

PyTorch : 75

75

Web Technologies

HTML/CSS : 70

70

PHP : 75

75

JS [JQuery, angular.js] : 80

80

Languages

C : 70

70

C++ : 50

50

Java : 80

80

python : 90

90

PROLOG : 60

60

DBMS

SQL : 80

80

MongoDB : 60

60

Tools and Frameworks

Weka : 50

50

Hadoop : 70

70

Django : 75

75

Owncloud : 65

65

Spring : 45

45

Ruby on rails : 45

45

Education

Stony Brook University

  • College of Engineering and Applied Sciences
  • Master of Science
  • add
  • Computer Science
  • 3.92/4

language
University of Mumbai

  • Sardar Patel Institute of Technology
  • Bachelor of Engineering
  • add
  • Computer Engineering
  • 8.76/10

work
Thakur College of Science and Commerce

  • High School
  • Bachelor of Engineering
  • add
  • Computer Science
  • 89.17%

person

Projects

  • 1    Live Soccer Match Prediction    [ Oct 2017 - Dec 2017 ]
    arrow_forward
    • Predicts the probability of each possible outcome at any point during the game.

    • Utilizing the initial betting odds as a Prior and the current game state to calculate the probability.

    • Converted a Time Series Prediction to a Classification problem. Also devised a new error metric [Linear offset Error].

    • Technology used - Python, Scikit-learn.

  • 2    Text-to-Image Synthesis    [ Nov 2017 - Dec 2017 ]
    arrow_forward
    • Implementation​ ​of​​ the​ paper​ - Generative​ ​Adversarial​​ Text​ ​to​ Image​ Synthesis [https://arxiv.org/abs/1605.05396].

    • Built on top of Deep Convolutional Generative Adversarial Network ​[DCGAN] to add text embedding features to the network.

    • Technology​ used - ​Python,​ PyTorch

  • 3    DNS packet Injector/Detector    [ Nov 2017 ]
    arrow_forward
    • A python script to test the DNS packet injection attack

    • Technology used - python, scapy

  • 4    Boosted SSH    [ Oct 2017 ]
    arrow_forward
    • Adding an extra layer of security to publicly accessible SSH servers.

    • Uses the ProxyCommand option of SSH

    • Technology used - C, OpenSSL, pthreads

  • 5    Phenotypic Prediction using transcriptomic features    [ Oct 2017 - Dec 2017 ]
    arrow_forward
    • Predicting Population origin and Sequencing center from a given output of Salmon.

    • Performed Feature Engineering to derive features from Equivalance Classes of transcripts.

    • Used feature selection by calculating the importance of each feature and then used thresholding.

    • Devised a technique for memory optimization owing to the large scale data.

    • Achieved an F-Score of 0.89 by using the MLP Classifier.

    • Technologies and concepts used - Python, Scikit-learn, Pandas, Multi-task prediction.

  • 6    Shakespeare style poem Generator    [ Jul 2017 - Jul 2017 ]
    arrow_forward
    • This Application uses Recurrent Neural Networks trained on poems of Shakespeare.

    • Can be used to generate a poem given a seed word.

    • Built two models - Character-based and word-based, the former generated better results.

    • Technology used - Python, NLTK, Recurrent Neural Network

  • 7    Google SVHN recognition    [ Jun 2017 - Jun 2017 ]
    arrow_forward
    • Implemented Deep Convolutional Neural Network using TensorFlow to recognize the house numbers in Street View House Number dataset.

    • Used Matlab to preprocess the images (i.e. to find the house numbers in entire images).

    • Technology used - Python, TensorFlow

  • 8    OpenBCI motion simulation    [ Jun 2016 - Apr 2017 ]
    arrow_forward
    • Built an application that classifies the direction of motion based on the EEG waves emitted by the brain using python and scikit-learn.

    • Can be used to move wheelchairs for physically challenged people.

    • Used Power Spectral density for feature extraction and PCA for dimensionality reduction of features.

    • OpenBCI system was used to record the EEG waves, laplacian and bandpass filters were used to preprocess the signal.

    • Technology and Concepts - Python, Scikit-learn, OpenBCI, Laplacian filter, Power Spectral Density, PCA

  • 9    Large Scale Bloom Filter using MapReduce    [ Mar 2017 - Mar 2017 ]
    arrow_forward
    • Built to check the number of unique customers for an online retail store.

    • Used FNV and Murmur hashing functions, to map the customer names to the bloom filter array.

    • Technology used - Java, Hadoop

  • 10    OCR    [ Dec 2016 - Dec 2016 ]
    arrow_forward
    • Devised an OCR system that uses Zoning technique to extract features from the segmented images of texts.

    • Implemented using MLPNN without optimizations

    • Technology used - Python and Scikit-learn.

  • 11    Behavioral finance project    [ Feb 2016 - Mar 2016 ]
    arrow_forward
    • Built an application that decides the best mutual fund policy for a given user on basis of multiple parameters

    • Used methods like Simple Additive Weighting (SAW), Weighted Product Model (WPM) and Analytical Hierarchy Process (AHP).

Contact

Email jbhatt@cs.stonybrook.edu
Alternate jay.yogeshkumar.bhatt@gmail.com