Summary
Overview
Work History
Education
Skills
Personal Blog on Machine learning
Other Projects
Accomplishments
Timeline
Generic

Arshad Kaji (Kazi)

Data Scientist

Summary

I'm a Data Scientist with deep interest in research based work and I have ambition to utilize skills and abilities and add valuable assets to an esteemed organization while being resourceful, innovative and flexible.

I have good communication and written skills to deliver the need of data expandability for businesses.

Overview

2
2
years of professional experience
4
4
years of post-secondary education

Work History

Associate Data Scientist

Morningstar
Mumbai
01.2021 - Current

I was responsible for the various machine learning and automation projects. By collaborating with multiple tech and non-tech teams, I have single handedly implemented many end to end deployments.


Smart Search (NLP, CV, Automation): A question-answering model for business documents. Using Huggingface transformer models, created an end to end project which satisfied business needs.

  • This project achieved 60% automation
  • Used huggingface Roberta Model for question answering
  • Created and deployed an API on Amazon AWS (Lambda)
  • Custom data preparation from HTML docs using Tesseract and regex logic


Document Classification (NLP, Automation) : To classify documents based the keywords and their values in the documents.

  • The deployed model achieved 92% accuracy with recall 99%
  • For various markets used SVM, Random Forest and XGBoost models
  • performed dataset preparation and data cleaning
  • achieved 100% automation , beating manual classification


Ripken Data classification (Tabular, Automation) : To automate the process of filtering junk data from the larger dataset.

  • Using great_expectations library created multiple expectations (filters)
  • Used various statistical techniques like normal curves and standard deviations for the data cleaning

Education

Bachelor of Engineering - Computer Science

Ramrao Adik Institute of Technology
Mumbai
06.2016 - 10.2020

Skills

    Keras /Tensorflow

undefined

Personal Blog on Machine learning

https://www.arshad-kazi.com

  • I write about mathematics, statistics behind various machine learning algorithms.
  • Deep learning Projects with stepwise explanations.
  • Various Deep learning neural nets and their architectures

Other Projects

Automatic Music Transcription (CV, Sequential):

  • An end to end deep learning-based model generates a musical script for the piano music.
  • The spectrograms of MIDIs are analysed using stacked LSTMs and CNNs models.
  • Implementation of a music transcription research paper. Libraries used: Tensorflow, Timidity, Scipy, sklearn.

Image Super-Resolution (CV):

  • A research based project which generates high resolution images for the blurry and low resolution images using deep learning model.
  • Research project in IIT Bombay.
  • Achieved a descent images using U-shaped CNN architecture.

Sign Language Recognition (CV):

  • Recognition of American Sign Language using CNN model.
  • More than 60 stars and forks on Github.
  • More than 94% test accuracy.

Accomplishments

    Among Top 3 in HackSRM, Amaravati

  • Implemented a CNN model for Heart Disease prediction.
  • Achieved 94% accuracy on the final solution.
  • Achieved best design and 2nd position award.


    • Coding Competitions & Other

  • Among Top 1% in CodeVita: A coding competition organized by TCS
  • Among Top 1% in HackWithInfy: A coding competition organized by Infosys
  • Gold badge in Python & Problem solving on Hackerrank.


Timeline

Associate Data Scientist

Morningstar
01.2021 - Current

Bachelor of Engineering - Computer Science

Ramrao Adik Institute of Technology
06.2016 - 10.2020

Similar Profiles

CREATE PROFILE
Arshad Kaji (Kazi)Data Scientist