Close

Shrobon Biswas

I'm

Download Resume Projects

Experience

Veeva Systems Inc(Toronto,Canada)

Role: Software Development Engineer
Team: Core-Renditions API and Infrastructure

Description: At Veeva, I have an opportunitiy of wearing multiple hats.

  • I mainly focus on feature/API development for core-renditions team
  • I am also responsible for anything Infrastructure or AWS related
  • Skills: Java(Spring), AWS (S3, SQS, Lambda, EC2, SQS, Step Functions), Terraform, Git

Amazon.com (Vancouver,Canada)

Role: Software Development Engineer
Team: Amazon Corporate Gift Cards

Description

  • Added 5 API's providing different reporting capabilities for AGCOD customers
  • Implemented API(s) are currently running in production environment of North America, Europe, China and Japan.
  • Engineered a scalable & timezone-adaptive scheduling mechanism for customised notifications
  • Skills: Java(Spring), JavaScript, AWS (S3, Lambda, Cloudwatch, SQS, DynamoDB), Continuous Deployment pipelines, JQuery, JUnit, PowerMock, EasyMock, Git

Attify Mobile Security Pvt.Ltd (Bengaluru, India)

Position: IOT Research Intern

Education

University of Alberta (Edmonton,Canada)

September 2016 - April 2018

Master of Science in Computer Science (Multimedia Specialization)
GPA : 4.0

Relevant Coursework

West Bengal University of Technology (Kolkata,India)

August 2012 - July 2016

Bachelor of Technology in Computer Science and Engineering
GPA : 8.83 / 10

Relevant Coursework

Skills

Projects

XYZUrl.me- URL Tracking & shortening

(WebApp + Chrome Extension)
With XYZurl users can shorten any mammoth sized url to a few characters only. API's have been created to also track the shortened URL's for visitor information like geolocation, ip, time of visit, browser etc, thus making it easier for businesses to understand their target audience.

  • Web app: http://www.xyzurl.me/
  • Chrome Webstore link: https://lnkd.in/eAc-gPV
  • Skills: NodeJs, Redis, MongoDB, Facebook login API, Javascript, Heroku-Cli, Git
View Project

Tweenalyzer- The Twitter Analyzer

Tweenalyzer helps analyze your twitter fanbase by providing a visually pleasing data-centric summary on different aspects of your target audience.

  • Twitter REST API was used to fetch the tweets based on the searh term
  • Python-Flask Framework was used for the backend
  • Data analysis was performed using scikit-learn and pandas library.
  • Google charts was used plotting the graphs on the keen.io dashboard.
View Project

Chinese Herb Classification using K-Means & Random Forest Classifier

  • This project achieved a classification acccuracy of ~98%
  • A python script was written to automate the process of herb extraction & labelling from given images.
  • K-means clustering was used to air the segmentation process.
  • Random Forest classifier was used train the extracted herbs
View Project

Convolutional Neural Networks to Classify Cats & Dogs

  • Increased classification accuracy to 91.18%
  • Utilized Random Search algorithm as a base for effective hyperparameter tuning
  • Performed just-in-time Image Augmentation, to tackle overfitting on the reduced Cats Vs Dogs dataset.
  • Stabilized the network by experimenting with regularizers, batch-normalization, dropouts & weight-initializations
View Project

Safari Rush Solver [AI Bot]

A fast AI bot to solve a Safari RushHour game

  • Uses SMA*(Simplified Memory-Bounded A* algorithm) & blocking vehicle heuristic
  • Given any board configuration, this solver is able to find a path with a minimum number of moves, to move the Safari Rover out of the board.
  • Increased solution search speed by 30% by using efficient data structures
  • This solver can find a solution of a moderately difficult board in around 10 seconds, while difficult and fiendishly hard boards take around 35 seconds.
View Project

Ataxx Solver [AI Bot]

An AI which can play the game of Ataxx optimally

  • Uses Alpha-beta pruning algorithm with iterative deepening.
  • Hashtables are used for memoizations of previous calculated results.
  • AI can be configured, to limit the search depth or limit time taken per move.
View Project

Cluster-Based Algorithm to count number of vehicles from Traffic Video

View Project

Publications / Certifications

Released a Video Tutorial Series on Computer Vision using OpenCV [Instructor]

Title: Hands on Computer Vision with OpenCV & Python

Sept 2016

IoT-MQTT based Energy Efficient Economic Home Automation System [first author]

9 July 2016

3rd International Conference on Microelectronics, Circuits and Systems, Micro2016.

Get in Touch