Soumi Bardhan
|
8697581335
|
https://github.com/Soumi7
|
|
https://twitter.com/BardhanMs
|
|
|
https://soumibardhan10.medium.com/
|
Kolkata, India
|
Education
Indian Institute of Information Technology, Kalyani
2018 - 2022
Bachelor of Technology in Computer Science
Current SGPA : 9.18/10
Work experience
GEEKs for geeks
Jan2021 - Mar 2021
Technical Writer Intern
Written several technical articles for coding and using python libraries. https://auth.geeksforgeeks.org/user/soumibardhan10/articles
Google Season of Docs 2020, SYMPY
Aug 2020 - Dec 2020
Technical Writer
Sympy is a lighweight independant computer algebra system.
Added and updated documentation, added missing style sections like Parameters. https://github.com/sympy/sympy/wiki/GSoD-2020-Report-Soumi-Bardhan:-Consistency-across-docstrings---SymPy-Documentation
Added and updated documentation, added missing style sections like Parameters. https://github.com/sympy/sympy/wiki/GSoD-2020-Report-Soumi-Bardhan:-Consistency-across-docstrings---SymPy-Documentation
Hiranya Growth Partners LLP
June 2020 - August 2020
Software Developer Intern
Part of the development team for Artificial intelligence in Finance.
Worked on Named Entity Recognition project for intent recognition from financial statements.
Used Flask, FastAPI, Docker and AWS to host the RESTAPI on cloud.
Worked on Named Entity Recognition project for intent recognition from financial statements.
Used Flask, FastAPI, Docker and AWS to host the RESTAPI on cloud.
ITMR Lab, IIIT Allahabad
May 2020 - July 2020
Computer Vision Research Intern
Research area included Self Simulatory Behaviours Analysis from video.
The work aims at early detection of Autism Spectrum Disorder in children.
The work aims at early detection of Autism Spectrum Disorder in children.
CPPSECRETS.com
Sept 2019 - Dec 2019
Python Developer
Worked on a pharmacy inventory management web application using Flask.
TECHNICAL ARTICLES
I always document projects as articles on Medium so that people can benefit from my experience. https://soumibardhan10.medium.com/
Named Entity Recognition — Simple Transformers —Flask REST API https://bit.ly/34EMMQG
Dockerizing and Hosting your Flask Web App(Rest API) on AWS EC2 https://bit.ly/3jNgrxj
Weather Based Stock Prediction with Pycaret
https://bit.ly/33I8MdZ
https://bit.ly/33I8MdZ
Achievements
Ernst AND YOUNG GDS
Aug 2020 - NOW
Finalist
One of the top 5 teams in the Natural Language Processing Domain. Developed a intent based file management application for easy file management for EY sharepoint repositories.
Developer Student Club, IIIT kalyani
Aug 2020 - NOW
Founder and Lead
Leading a community of students to develop innovative solutions to solve problems faced by local businesses in my university. Enabling students at our institute to have their projects featured on global platforms and learn new technologies.
NETAPP WIN FINALIST
-
Hack404 - Team Lead
One of the 17 teams in NetAPP WIN Finals selected for two month mentorship by Netapp experts. Presented my healthcare solution - a mobile app detecting poses in real time which implements on-device machine learning.
P&G Global INNOVATIOn CHallenge FINALIST
-
Team Lead
One of the 8 teams in P&G Global innovation challenge. We used neural Networks and SVM to create a model for Weather based Stock Prediction On Walmart FMCG Retail dataset.
ODSC West diversity scholarship
-
Got selected to receive an academic pass to ODSC West Conference 2020
MICROSOFT AZURE ML SCHOLAR
Aug 2020 - NOW
Was selected by Udacity and Microsoft to receive a scholarship for Microsft Azure Machine Learning course.
Skills
Python, Java, C++, C
Advanced
Tensorflow, Pytorch
Intermediate
Data Structures and Algorithms
Advanced
Docker
Intermediate
AWS, Microsoft Azure
Advanced
Flask, Django, HTML, CSS, JS
Advanced
Projects
Sofia - AI Powered Yoga Instructor
Sofia is an artificially Intelligent yoga trainer flutter app which can detect yoga poses in real time. Uses tensorflow-flite for ON-Device ML. The yoga pose dataset was obtained and cleaned manually. The model used transfer learning on VGG16. The f1 score achieved was 0.98. The model is hosted online with AWS and has REST API endpoints for external access. The flutter app uses these endpoints to send snapshots to the model and get the pose as output. Uses tensorflow-flite for ON-Device ML to remove the latency issue for multiple simultaneous users. https://github.com/sbis04/yoga-instructor.
Named Entity Recognition
Named-entity recognition (NER) is a subtask of information extraction that seeks to locate and classify named entities mentioned in unstructured text into pre-defined categories. My project included NER from financial data corpus and performing intent recognition from the sentences as well, which was hosted as a web application online using flask for backend. The NER Model used simple transformers library to train the data. The f1 score achieved was 0.87 for fine tuning on roberta-base and 0.92 for roberta-large.
Autism Spectrum Disorder Detection
Detection of Autism Spectrum disorder by self simulatory behaviours analysis through video and image processing.
FMCG Stock management for P&G
Uses Neural Networks and SVM to create a model for Weather based Stock Prediction On Walmart FMCG Retail dataset. For the respective dates, the data was obtained using the wwo-hist API. Several regression models we trained with pycaret. Light gradient boosting machine gave the lowest RMSE. https://github.com/Soumi7/PG-StockManagement.
Codery - Zulip Bot for Competitive programming
A bot which integrates several API s with the zulip chat platform. It also scrapes data from several websites to display results directly into zulip chat. https://github.com/Soumi7/codery
Sofia is an artificially Intelligent yoga trainer flutter app which can detect yoga poses in real time. Uses tensorflow-flite for ON-Device ML. The yoga pose dataset was obtained and cleaned manually. The model used transfer learning on VGG16. The f1 score achieved was 0.98. The model is hosted online with AWS and has REST API endpoints for external access. The flutter app uses these endpoints to send snapshots to the model and get the pose as output. Uses tensorflow-flite for ON-Device ML to remove the latency issue for multiple simultaneous users. https://github.com/sbis04/yoga-instructor.
Named Entity Recognition
Named-entity recognition (NER) is a subtask of information extraction that seeks to locate and classify named entities mentioned in unstructured text into pre-defined categories. My project included NER from financial data corpus and performing intent recognition from the sentences as well, which was hosted as a web application online using flask for backend. The NER Model used simple transformers library to train the data. The f1 score achieved was 0.87 for fine tuning on roberta-base and 0.92 for roberta-large.
Autism Spectrum Disorder Detection
Detection of Autism Spectrum disorder by self simulatory behaviours analysis through video and image processing.
FMCG Stock management for P&G
Uses Neural Networks and SVM to create a model for Weather based Stock Prediction On Walmart FMCG Retail dataset. For the respective dates, the data was obtained using the wwo-hist API. Several regression models we trained with pycaret. Light gradient boosting machine gave the lowest RMSE. https://github.com/Soumi7/PG-StockManagement.
Codery - Zulip Bot for Competitive programming
A bot which integrates several API s with the zulip chat platform. It also scrapes data from several websites to display results directly into zulip chat. https://github.com/Soumi7/codery
CAREER Profile
DSC Lead IIITKalyani, involved in research in Computer Vision and ML domain. I am an enthusiastic open source contributor, I enjoy solving problems on leetcode and am have contributed to Sympy as a part of Google Season of Docs 2020.
