HELLO, I'M
Soumava Banerjee
About Me!
I am a graduate student pursuing my Masters in Computer Science and Engineering, at Ohio State University, Columbus, Ohio, USA. Before opting in to a foreign graduate program, I completed my Bachelors from Jadavpur University, India in Electronics & Telecommunication Engineering.
For my professional experience, I am currently working as a Machine Learning
Engineer Intern, after completing a successful two and a half year Software Development experience with Goldman Sachs. Besides I am also enrolled as a Graduate Research Associate with Ohio State University working as a Data Science specialist on a Ohio State initiative to collaborate researchers and funding partners.
As for leisure times,I like swimming and travelling a lot. I also like playing cricket and badminton and occasionally strum on my guitar.
Contact Me!
Email: soumavabanerjee0519@gmail.com
Phone: (+1) 614-619-1245
Location:
1466, Worthington St., Apt.-D, OH - 43201
Education
WHAT I’VE LEARNED
Aug. 2019 - Present
Ohio State University – Columbus, OH; Class of 2019
Master of Science; Computer Science & Engineering
GPA: 3.7/4.0
Experience
WHERE I’VE WORKED
May. 2020 – Aug. 2020
Machine Learning Engineer Intern, Astute Solutions
Dec. 2019 - Present
Ohio Innovation Exchange, Graduate Research Associate
Jul. 2013 – Aug. 2017
Jadavpur University – Kolkata, WB; Class of 2017
Bachelor of Engineering; Electronics & Telecommunication Engineering
GPA: 9.12/10.0
Apr. 2017 - Aug. 2019
Goldman Sachs Services Pvt. Ltd., Software Development Engineer
Apr. 2016 - Aug. 2016
Goldman Sachs Services Pvt. Ltd., Summer Analyst
Jan. 1998 - Jan. 2013
South Point School, Kolkata, WB Class of 2013
Percentile ~94
Projects
Here's take a look at some of the things I have done!
Classification of Customer Complaint Emails using Pre-trained Models
Applied Natural Language Processing to extract information out of customer complaint emails across various food chains of USA, to classify them according to their level of priority and actions, in order to support automation following COVID-19 outbreak. Used hugging face library of tensorflow and ktrain library to pre-train text records using various pre-trained weight models like BERT and fastText and ultimately use transformers for classifying emails both for single labels and multi-labels. Achieved an accuracy of ~65% outperforming other sophisticated model performances (Inception Network)
Part of summer research industrial internship [No open source code available hence]
Math Word Problem Solver
Scraped mathematical word problems from Yahoo Answers using an open source web-crawling tool called Scrapy. Filtered and augmented valid records to be used with a combination of deep learning models (Bi-LSTM Encoder Decoder Network with Attention Mechanism for the Equation Tree Generation and Recursive Neural Networks with a self-attention layer for the answer prediction) to generate equation templates for the question and predict the answer.
English Premier League Season Champion Prediction
Scraped official fifa websites to gather information on player stats and match fixtures of previous seasons to predict the outcome of current season matches. Used simple deep neural networks for training and prediction of results achieving an accuracy of ~67%
Designing Interpreter for a formal language : Quandary
Implemented dynamic and static type checking, garbage collection (Mark-Sweep) along with dynamic memory allocation (use of free keyword), full concurrent behavior with synchronized blocking functionality with a formal language named Quandary built with JAVA.
[github code link coming up]
Skills & Languages
WHAT I BRING TO THE TABLE
Languages
JAVA
C++
Python
Matlab
Software Development
Backend TechStack
-
Rest Services - Jax-RS (Jersey), Spring, Dropwizard, Swagger API, Asynchronous Rest Services with observables using Rx2Java, Flask
-
Database Connectivity - JDBC, Hibernate, JDBI3 using Dropwizard
-
Reactive micro-service architecture using Vertx io (polygot), Hazelcast for scaling, event driven across components with Apache Kafka for asynchronous streaming
-
Containerization using Docker
FrontEnd TechStack
-
Basic FrontEnd glossaries like HTML, Javascript, CSS
-
ReactJS framework with Redux as a state container
Other Items
-
Deployment using Subversion (trunk-based deployment) and git
-
Groovy, build systems for Java Gradle and Maven
-
Shell scripts for writing schedulers and adhoc searches.
-
Numpy, Pandas (dataframe and sqldf) Python
-
Elastic Search
Machine Learning
-
Email Classification (Single Label and MultiLabel) using Hugging Face (Transformers), KTrain, pre-trained models like fastText, Google BERT
-
NLTK, EDA for Data Augmentation
Awards
WHERE I SHINE
- Received Pre-Placement Offer for successfully converting an internship from Goldman Sachs
- Ranked 268 in WBJEE, 2013
- Received Graduate Research Associate-ship with Full Tution Fee Waiver and Stipend support for Research Contribution at Ohio State University
Interests
OUT OF OFFICE
Tennis
Swimming
Travel
Yoga
Reading