myAvatar
About
Education
Experience
Skills
Publications
Projects
Positions of Responsibility
Courses
Achievements
Resume

*Desktop compatible view only*

myAvatar

Manas Bundele

About

I am a recent graduate from The University of Texas at Dallas with MS in Computer Science and a specialization in Intelligent Systems. During the summer of 2020, I worked as a Machine Learning Intern at Etsy Inc. on the design and development of a novel technique towards a scene-aware, compatible fashion product recommendation system. I am always excited to talk about interesting problems in the field of Computer Science.


I completed my B.Tech. in Computer Science and Engineering
at Indian Institute of Technology, Jodhpur in 2016. After my B.Tech., I worked as a Full Stack Software Engineer at Voylla Fashions Pvt. Ltd. where I worked on several key projects using Ruby on Rails(backend), HTML/CSS/JS/JQuery(frontend) and Postgresql(database). I am a quick learner and a diligent individual with strong analytical and coding skills.


I am actively looking for full-time roles (application/research-based) as
Machine Learning Engineer, Software Engineer, Computer Vision Engineer and
Natural Language Processing Engineer.

Education

mail

University of Texas at Dallas

| MS in Computer Science (Thesis)


August 2020

- Recipient of Jonsson School Graduate Study Scholarship (2018-19)
- Recipient of Master's Research Fellowship (2020)

mail

Indian Institute of Technology, Jodhpur

| B.Tech. in Computer Science & Engineering


May 2016

- B.Tech. Project nominated for 'Best B.Tech. Project' award in the institute

Experience

mail

Etsy Inc.

| Machine Learning Intern

Jul 2020 - Oct 2020


Project: ‘Let Etsy Pick Your Shoe for You!’

  • Designed and implemented a novel technique towards scene-aware, compatible fashion product recommendation with over 5% improvement in accuracy over the chosen baseline on STL dataset.
  • Proposed approach consisted of a pre-trained Resnet50V2 model, along with specialized layers and KNN to learn the style compatibility between a user-uploaded outfit and a product from Etsy’s inventory (without fine-tuning) using Triplet Loss
    • python tensorflow scikit-learn


    mail

    UT Dallas Computer Science Department

    • Outreach Coding Camp Counselor

      • Jun 2019 - July 2019


    • Guided and mentored young impressionable minds during summer coding camps

    • Computer Science Grader

      • Aug 2018 - May 2019 | Feb - May 2020


      • Got nominated for the 'Best Grader' Award in the department
      • Graded the tests, quizzes and assignments of a batch of 75 undergraduate students for the course Discrete Mathematics



      mail

      Voylla Fashions Pvt. Ltd., Jaipur

      | Software Engineer

      May 2016 - Jun 2017


      • Project LP Boost: Designed and implemented a Landing Page Ranking algorithm that increased the Average Order Value by 30% in collaboration with the Marketing & Business Development team.
      • Project Solidus: Voylla Website Upgradation (Team of Eight developers) and Data Migration.
        - Developed and optimized order pipeline(in a team of 4), improving its performance (~seconds) for seamless experience.
        - Handled data migration, a very critical aspect of the project which could only be totally tested in production environment
        - Improved the efficiency of Order Pipeline and other features through code and database query optimization
      • Project CRM Centre Database: Wrote optimized Postgresql scripts to synchronize the customer’s data from website and ERP databases into a centralized database daily, data being used for analysis and in key projects by the Data Science team
        • ruby on rails postgresql html css jQuery javascript Git AWS solr

Skills

Below are some of the skills that I have used in my projects before. I am always willing to add more to my skillset.

I have been using Python extensively for over the past 3 years for my coursework and research. Most of my projects related to CV, NLP and ML have been coded in Python which can be found on my Github profile.

I have been using this library over the past year in my projects as well as my research work. I have also completed a course 'Tensorflow in Practice' by deeplearning.ai. I have extensively used it during my internship at Etsy to develop a recommendation system.

During my job at Voylla Fashions Pvt. Ltd., being one of the go-to backend developers, I have heavily used it for development, debugging and optimization of several projects. I have written several thousands of lines of code and also debugged several critical issues during this period.

I have used Scala in all the assignments during my course, Big Data Analytics and Management. It is a wonderful language to work with and hope to use it again in my other projects.

Since the summer of 2019, I have been using technologies in the Big Data ecosystem i.e. Hadoop, Mapreduce, Apache Spark and NoSQL systems.

I have extensively used C/C++ in projects and competitive coding during my B.Tech., prior to 2016. Recently, got in touch with it again(alongwith opengl) in my course Computer Animation and Gaming.

I have written over 5000 queries in Postgresql over the course of my projects at Voylla. From using it for analysis, data cleaning and data migration in Solidus to creating a CRM centre database(centralized database for ERP and Website), it has been an interesting journey working with data using Postgresql.

Right from building this website to working on frontend related projects at Voylla, it has been fun using HTML, CSS, JS and Jquery.

Publications


Projects

This list of projects cover most of my projects but is not limited to the same.
Click on a project for more details about it.

@University of Texas at Dallas

Caption Generation*

[Master's Thesis]

To develop a novel technique to generate the description of an image by understanding its contents. Proposed a novel technique for ranking different regions of the image on the basis of their saliency. Developed 3 novel techniques to discern insightful contexts by incorporating commonsense knowledge.

python nltk spacy lightgbm pyspark word embeddings knowledge graphs

Fake Opinion Detection

[Big Data]

Developed a supervised learning model using Pyspark and MLlib that was able to detect fake reviews by learning from YelpNYC dataset.

pyspark mllib nltk sklearn

Sentiment analysis of live
streaming twitter feeds

Created a Spark Streaming application that continuously reads data from Twitter about a topic and these feeds can be analyzed for their sentiment and the trend can be visualized using ElasticSearch, Logstash and Kibana(ELK) stack.

scala apache spark kafka
elasticsearch logstash kibana stanford corenlp sentiment analyzer

Other Big Data-related projects

Implemented various algorithms and performed analysis on big data related to airports, movies, twitter social circles, etc.

  • Pagerank for airports
  • Search Engine for movies
  • Topic modeling for classic books
  • Twitter airline sentiment analysis
  • Twitter social circles analysis using GraphX

scala apache spark hadoop mapreduce graphx mllib sparksql

Modified MNIST Digit Recognition
using ResNet

Implemented a residual network using Tensorflow in Python and included batch normalization, non-linearities, dropout, l2 regularization to improve accuracy and reduce overfitting. Achieved over 90% accuracy in 4 & 1/2 min on testing data of 10k images with modified different random training datasets of 6k images of size 7*7

python opencv tensorflow keras

Wink Detection using Haar Cascades

[Computer Vision]

Implemented a winking face detection algorithm in images as well as live video using Haar feature-based cascade classifiers.

python opencv

Text Generation using N-Gram Language Model

Implemented a MLE based N-Gram language model with various smoothing techinques to generate text based on a written text.

python

Sentiment Analysis of
Movie Reviews

Implemented a sentiment analyzer using Naive Bayes and Logistic Regression classifier to predict the sentiment of a review. It also uses a sentiment lexicon, Dictionary of Affect in Language, which is further augmented using WordNet to improve sentiment analysis results.

python sklearn nltk

POS tagging using MEMM
with Viterbi decoding

Implemented a POS tagger using Maximum Entropy Markov Model with Viterbi to decode the highest probability sequence of tags for each test sentence.

python sklearn nltk scipy

@IIT Jodhpur

Interactive Medical Image Segmentation

[B.Tech. Project]

Developed a semi-automated segmentation technique for efficient image analysis for 3D visualisation as well as image guided surgery at reduced costs and minimized expert intervention. Got nominated for the Best B.Tech. Project Award, one of the 3 projects nominated in the entire batch of 2016.

c++ opencv matlab

Interactive Medical Image Segmentation

[B.Tech. Project]

Link to Paper

Centralized Video Processing

Developed a remote video processing system using socket programming in C++. Multiple users were managed using threading. Reduced the network load and video tagging was done at the server for minimizing data storage.

c++ opencv

Virtual Memory Management

Designed a virtual memory system in C and analysed the performance of FIFO, clock and LRU algorithms for page replacement. Parameters considered – prepaging vs demand paging, local vs global page replacement, precleaning.

Gest-Talk

An application, developed within 24 hours, in C# during Code.fun.do(2015) conducted by Microsoft, using the machine learning feature of Visual Gesture Builder for Microsoft Kinect 2.0,which targets the communication between differently-abled people with the help of gestures.

c# visual gesture builder kinect

Positions of Responsibility

Below are some of the major PORs that I managed during my B.Tech.

Assistant Student Head, Counseling Service, IIT Jodhpur

Mar 2014 – Apr 2015


Led and managed a team of 17 student guides along with 7 other assistant student heads and successfully set up a Counseling Referral System and conducted Career Counseling Sessions for the 1st time in the institute.


Head, Publicity and Media, Varchas'14, IIT Jodhpur

Sep 2013 – Feb 2014


Led a two-tier team of 12 people to plan and publicize the inter-collegiate sports festival of the institute, bringing countrywide participation of over 1000 people. Publicized and brought a participation of over 500 people in the Varchas Mini-Marathon’14, organized across Jodhpur by IIT Jodhpur


Coordinator, Marketing, Ignus’14, IIT Jodhpur

Sep 2013 – Feb 2014


Worked with a team of 26 members to garner Rs.30 lakhs as sponsorship for the inter-collegiate techno-cultural festival of IIT Jodhpur, bringing a footfall of over 4000 people.


Student Guide, Counseling Service, IIT Jodhpur

Mar 2013 – Apr 2016


Mentored a group of 10 counselee students and guided them regarding academic and extra- curricular affairs in the institute, acting as a Student Guide throughout their stay.


Courses

@University of Texas at Dallas


  • Machine Learning
  • Computer Vision (A+)
  • Convolutional Neural Networks
  • Natural Language Processing
  • Design & Analysis of Computer Algorithms
  • Big Data Management And Analytics
  • Distributed Computing
  • Database Design
  • Computer Animation and Gaming
  • Artificial Intelligence

@IIT Jodhpur


  • Artificial Intelligence
  • Digital Image Analysis
  • Computer Graphics
  • Operating Systems
  • Compiler Design
  • Computer Networks
  • Principles of Databases Systems
  • Software Engineering

Achievements

Few of the many things to be proud of.
  • Recipient of Master’s Summer Research Fellowship at UTD. (2020)
  • Recipient of Jonsson School $1000 Graduate Study Scholarship at University of Texas at Dallas (2018-19)
  • Journal Publication honored with "Meritorious" status in Honored papers 2017, Computers in Biology and Medicine (2018), doi: 10.1016/j.compbiomed.2018.05.020
  • B.Tech. Project, one of the three projects nominated from the institute for 'Best B.Tech. Project' award (2016)
  • Among the top 0.7% students in the country; securing All India Rank 3471 in IIT-JEE 2012 among 500,000 candidates
  • Secured State Rank 30 (2009) and District Rank 1 (2008) in Maharashtra Talent Search Examination
  • Stood among the top 0.01% students in the Maharashta State Scholarship examination in 2007

A bit more about me...

Although the grad life keeps me busy nowadays, I try to invest time in my new-found love for cooking new delicacies and fitness. I also enjoy playing basketball. I love drawing and painting and have even surprised myself by making a few good ones. I also love getting my face painted though. :D

Resume