Hi, I am Maitreyee Mhasakar

Geek. Engineer. Coder. Machine Learning Enthusiast.


Team Darth Vader. Team White Walker.

Know more

About Me

profile picture

My passion for coding stems from the core fun of logic building. Whether it was Cloud Computing five years ago, Machine Learning today, or Quantum Computing in the future; you will always find me building algorithms to turn ideas into real-world applications.

I have a pending patent for a Pattern Recognition system and two solid internships of automating manual decision-making processes using Machine Learning and Natural Language Processing that reduced human effort from hours/days to minutes. At Pal Inc, I built ML models for early meltdown predictions in Austism and spearheaded effort of data migration to AWS.Currently, I'm working at Lumenci Inc. where I help law firms understand technology in Software litigation cases and help software companies in Intellectual Property Monetization.

I'm on the lookout for an exciting full-time opportunity where I get to challenge myself, utilize my skill set to contribute to path-breaking innovation, and drive projects to success. I look forward to creating my niche in tech, contributing to impactful projects, and making a difference. I'm glad you are here, explore my site to see all that I have to offer.

Resume

Skills

Python

75%

Java

50%

C++

40%

NLTK

60%

Spacy

50%

Scikit-learn

50%

Numpy

50%

Pandas

60%

Tensorflow-Keras

30%

PyTorch

20%

Microsoft Azure AI

40%

MySQL

40%

PostgreSQL

30%

MongoDB

30%

HTML-CSS

60%

Javascript-Reactjs

30%

PHP

40%

OpenSource

30%

JIRA

30%

GitHub

30%

Jupyter,Google Colab

70%

Education

University
   


University of Texas at Dallas  Aug 2018 - May 2020

Masters in Computer Science

3.4/4.0

Machine learning, Algorithms, Natural language processing, Software evolution and maintenance, Artificial Intelligence, Web Programming Languages

University
   
College


Savitribai Phule Pune University  Jul 2013 - Jun 2017

MKSSS Cummins College of Engineering for Women

Bachelors in Computer Engineering

74%, First Class Distinction

Data Strcutures, Operating Systems, Database Management, Cloud Computing, Smart Systems

Work Experience

Associate   June 2021 - Present

Lumenci Inc


Work as technical consultant to provide diverse technology related insights to software companies and law firms in Intellectual Property litigation cases.



Data Science Intern   Aug 2020 - May 2021

Pal Inc (Gaia Wearables)


Building Machine Learning model on wearables’ sensor data for predicting early meltdowns in Autism Spectrum Disorders.



Machine Learning Engineer Intern   Jun 2019 - Aug 2019

DPR Construction


  • Automated end-to-end manual sentiment analysis process with 74% accuracy.
  • Developed customized NLP, ML modules to facilitate supervised classification of construction industry data.
  • The designed system improved manual project selection decision-making procedures by 70%.


Research Assistant   Jul 2017 - Apr 2018

MKSSS Cummins College of Engineering for Women


  • Published Patent as one of the three inventors working for Cummins College of Engineering for Women.
  • Designed an innovative handwritten Devanagari character classification algorithm based on a variant of density-based clustering system that reduced processing time of existing systems from complexity of n^2 to nlogn & optimized space usage complexity from n2 to n with 70% accuracy.


Project Intern (Software Developer Intern)   Aug 2016 - May 2017

Principal Financial Group


  • Worked in a team of 4 to automate existing manual underwriting process with 80% accuracy.
  • Built an end-to-end system that performed extractive summarization of medical documents using regex and Natural Language Processing, explored and harnessed Machine Learning algorithms for classifying client into appropriate risk class.


Projects

Patent Publication: Handwritten Devanagari Character Recognition System

  • My native language is Marathi and we write it using Devanagari script.
  • Very few systems have successfully digitized handwritten documents due to complexity of its structure and variability.
  • I designed an efficient new character recognition algorithm that identifies characters based on structure and not on size.
  • This will help revive old scriptures and documents, digitize daily local transactions and help bring tech to people who speak and use the native language in the region as well help bring it to the world.



See Publication



System to score meaning based sentence similarity

  • Built a Natural Language Processing based semantic system to score English sentences based on meaning similarity.
  • Predicted score in range of 1-5; 1 as least similar and 5 as most similar with 70% accuracy.



Source Code



French to English Language Translation

Built a system to translate 100,000 French sentences to English with 87% accuracy.




Source Code



Clothing Classification using Deep Learning

Built a convolutional neural network on a popular Fashion MNIST dataset of 60,000 clothing images to classify clothing into one of 9 categories with 94% accuracy.




Source Code



Paralleling tuning of PID controller simulations using GPU

  • Converted serial operation of PID controller simulation tuning into parallel by harnessing GPU computing power using CUDA on Nvidia GTX 1060.
  • This reduces processing time and improves computation with accuracy of 85%.






Virtual Reality based Chemistry lab

  • Built a simulation of a chemistry lab where user performs experiments using various physics and object effects on the apparatus and interacts using navigation techniques, throwing, grabbing and other body movements.
  • The simulation can help chemist analyze experiments before the actual trials in the lab.
  • This can help teach apparatus handling to Chemistry beginners.






Contact

Would you like to work with me? Awesome!

maitreyee.mhasakar@gmail.com