Career Profile
Pursuing Masters in Computer Science and been a Graduate Research Assistant since 1 year in the field of Reinforcement Learning and majorly contributed in a journal publication, named 'Team Learning from Human Demonstration with Coordination Confidence'.
A Technology Analyst with 2 years of experience in Analysis, Design, Development, Bug fixes, Support and Implementation of various stand-alone, client-server enterprise applications with 1+ years hands-on experience in ETL using Informatica Power Center 9.5, writing complex SQL queries and PL/SQL scripts. Contributed to building web pages using HTML and CSS.
Experiences
- Worked on a paper named 'Team Learning from Human Demonstration with Coordination Confidence Environment' in Python environment and submitted as a co-author
- Researched information regarding Reinforcement Learning to assist professors with academic pursuits and worked on developing a technique to improve the bootstrapping methods for training reinforcement learners in team-based collaborative environments.
- Experimented and studied the learning performance of Tabular Q-Learning techniques with different types of bootstrapping methods for learners with human demonstrations.
- Developed a team-based game domain named Guided Navigation (two variants) using Python Gym libraries.
- Planned, modified, and executed research techniques, procedures, and tests.
Thesis - Team Learning from Human Demonstration with Coordination Confidence
Environment: Python, Ubuntu OS, OpenAI Gym libraries
Description: A research work involving study, development and experimenting with reinforcement learning techniques and methods of bootstrapping them using prior human knowledge. These methods are experimented on custom made game domains with two agents using OpenAI Gym libraries.
Responsibilities: - Responsible for setting up the environment required to test the learning techniques, i.e. setting up python virtual environment, gym environment (game domains for RL testing) and Baselines (one of the standard open-sources for RL learning techniques). - Developed a team-based game domain named Guided Navigation (two variants) using Gym libraries. - Worked with tabular Q-Learning techniques to observe the learning performance of various techniques using human demonstration to bootstrap the learner. - Worked in collaboration with the faculty advisor in completing a paper named 'Team Learning from Human Demonstration with Coordination Confidence Environment' in Python environment.
Description: A research work involving study, development and experimenting with reinforcement learning techniques and methods of bootstrapping them using prior human knowledge. These methods are experimented on custom made game domains with two agents using OpenAI Gym libraries.
Responsibilities: - Responsible for setting up the environment required to test the learning techniques, i.e. setting up python virtual environment, gym environment (game domains for RL testing) and Baselines (one of the standard open-sources for RL learning techniques). - Developed a team-based game domain named Guided Navigation (two variants) using Gym libraries. - Worked with tabular Q-Learning techniques to observe the learning performance of various techniques using human demonstration to bootstrap the learner. - Worked in collaboration with the faculty advisor in completing a paper named 'Team Learning from Human Demonstration with Coordination Confidence Environment' in Python environment.
- Collaborated with clients from concept through final delivery of product or service.
- Coordinated with systems partners to finalize designs and confirm requirements.
- Provided continued maintenance and development of bug fixes and patch sets for existing web applications.
- Proposed technical feasibility solutions for new functional designs and suggested options for performance improvement of technical objects.
- Extensively involved in development of applications and processes using Informatica Power Center and Python framework.
- Developed PL/SQL procedures for Informatica mappings to implement run-time DQM (Data Quality Management) checks.
- Involved in various testing phases like Unit Testing, System Integration Testing and Performance Testing and migrated the code into QA (Testing) and supported QA team and UAT Users.
- Extensively involved in report generation using MS Excel data analysis and Client Report generation.
Project 1 - Abbvie Roster Manager
Environment: Javelin Application Catalyst based on Python Framework
Description: A client specific customized online application which is used to track all the medical representatives of AbbVie and their assignments to the Prescribers across various territories. Worked on enhancements of the existing application.
Responsibilities: - Gathered technical level requirements from clients. - Developed enhancements on the tool using Javelin specific python framework. - Created test cases to test various negative and positive scenarios. - Involved in performance tuning of the tool.
Description: A client specific customized online application which is used to track all the medical representatives of AbbVie and their assignments to the Prescribers across various territories. Worked on enhancements of the existing application.
Responsibilities: - Gathered technical level requirements from clients. - Developed enhancements on the tool using Javelin specific python framework. - Created test cases to test various negative and positive scenarios. - Involved in performance tuning of the tool.
Project 2 - Abbvie Business Rule Manager
Environment: Python, Informatica, Teradata, SQL
Description: A client specific application which is a combination of customized online application used by the clients to define the business rules and Informatica process chain used to process the Medical Representatives’ and Prescriber data based on the defined rules on UI. Worked on UI enhancements, existing Informatica process improvements and new Informatica processes.
Responsibilities: - Gathered technical level requirements from clients. - Developed enhancements on the tool using Javelin specific python framework. - Developed Informatica mappings to process data. - Created test cases to test various negative and positive scenarios for both UI and Informatica processes. - Involved in performance tuning at SQL level and UI level.
Description: A client specific application which is a combination of customized online application used by the clients to define the business rules and Informatica process chain used to process the Medical Representatives’ and Prescriber data based on the defined rules on UI. Worked on UI enhancements, existing Informatica process improvements and new Informatica processes.
Responsibilities: - Gathered technical level requirements from clients. - Developed enhancements on the tool using Javelin specific python framework. - Developed Informatica mappings to process data. - Created test cases to test various negative and positive scenarios for both UI and Informatica processes. - Involved in performance tuning at SQL level and UI level.
Project 3 - Abbvie Account and Targeting Feedback Management Tool
Environment: Python
Description: A client specific customized online application which is a supplement to Business Rules Manager and is only for HQ people to make and request changes to the business rules on territory basis for every quarter. Worked on complete development of the project release phase-one.
Responsibilities: - Gathered technical level requirements from clients. - Developed enhancements on the tool using Javelin specific python framework. - Created test cases to test various negative and positive scenarios. - Involved in performance tuning of the tool.
Description: A client specific customized online application which is a supplement to Business Rules Manager and is only for HQ people to make and request changes to the business rules on territory basis for every quarter. Worked on complete development of the project release phase-one.
Responsibilities: - Gathered technical level requirements from clients. - Developed enhancements on the tool using Javelin specific python framework. - Created test cases to test various negative and positive scenarios. - Involved in performance tuning of the tool.
Projects
Gym Two-Agent Game Domains
Environment: Python, Ubuntu OS, OpenAI Gym libraries
Description: A two-agent game domain with goal and reward orientation developed using OpenAI Gym libraries
Responsibilities: - Design the game domain and layout of the environment - Analyze the observations and actions for the agents - Analyze the state space of the environment - Develop the game environment and test the best implementation of state, action and observation spaces.
Description: A two-agent game domain with goal and reward orientation developed using OpenAI Gym libraries
Responsibilities: - Design the game domain and layout of the environment - Analyze the observations and actions for the agents - Analyze the state space of the environment - Develop the game environment and test the best implementation of state, action and observation spaces.
Tabular Q - Learning Techniques
Environment: Python
Description: Tabular Q - Learning combined with Reinforcement Learning Techniques to train multi-agent game domains
Responsibilities: - Develop the algorithm for two-agent game domains - Develop learning menthod calls using different bootstrapping methods.
Description: Tabular Q - Learning combined with Reinforcement Learning Techniques to train multi-agent game domains
Responsibilities: - Develop the algorithm for two-agent game domains - Develop learning menthod calls using different bootstrapping methods.
NoSQL Databases - Performance Comparison
Environment: Express, NodeJS, Cassandra
Description: A performance comparison project between Cassandra and MongoDB for specific transactions over a warehouse. The warehouse is of drug information with very huge database generated with sample online data.
Responsibilities: - Setting up the warehouse on Cassandra and MongoDB (including schema design, node setup on Cassandra and data population). - UI design and development. - Application setup and performance measure.
Description: A performance comparison project between Cassandra and MongoDB for specific transactions over a warehouse. The warehouse is of drug information with very huge database generated with sample online data.
Responsibilities: - Setting up the warehouse on Cassandra and MongoDB (including schema design, node setup on Cassandra and data population). - UI design and development. - Application setup and performance measure.
Oracle 12C – Distributed Database Setup Feasibility Check
Environment: Oracle Database 12c, Docker
Description: A feasibility check of real-time communication between instances of Oracle 12c databases with similar schemas. University database instances of two different campuses are set at different locations which have the same online user interface.
Responsibilities: - Setup different database instances using Oracle DB images on docker - Create the University Schema on both the instances - Setup communication between the instances
Description: A feasibility check of real-time communication between instances of Oracle 12c databases with similar schemas. University database instances of two different campuses are set at different locations which have the same online user interface.
Responsibilities: - Setup different database instances using Oracle DB images on docker - Create the University Schema on both the instances - Setup communication between the instances
Site Performance Check by Site Scraping
Environment: Express, NodeJS, Cassandra
Description: A web-scraping tool for specific site to perform automated user actions and collect data overtime, to check the load time performance of the site overtime. Cassandra database is used to effectively store the time-series data and retrieve the information later to plot the performance analysis.
Responsibilities: - Develop web-scraping tool for the site. - Setup a Cassandra database for storing temporal data from user actions. - Automate user actions and collect data. - Plot the load time performance of the site using collected data.
Description: A web-scraping tool for specific site to perform automated user actions and collect data overtime, to check the load time performance of the site overtime. Cassandra database is used to effectively store the time-series data and retrieve the information later to plot the performance analysis.
Responsibilities: - Develop web-scraping tool for the site. - Setup a Cassandra database for storing temporal data from user actions. - Automate user actions and collect data. - Plot the load time performance of the site using collected data.
Online Car Rental System
Environment: Flask, MySQL
Description: A simple online software solution for car rental enterprises which allows the users to perform actions like car bookings, cancellations, etc. using Python Flask and MySQL database.
Responsibilities: - Setting up MySQL database (including schema design and sample data population). - UI design and development using Flask framework. - Application setup and implementation / testing of various use-cases.
Description: A simple online software solution for car rental enterprises which allows the users to perform actions like car bookings, cancellations, etc. using Python Flask and MySQL database.
Responsibilities: - Setting up MySQL database (including schema design and sample data population). - UI design and development using Flask framework. - Application setup and implementation / testing of various use-cases.
Fuzzy Based Edge Detection of Grayscale Images
Environment: Java and Eclipse IDE
Description: A project involving study and development of an algorithm to detect edges of grayscale images using first derivative algorithms and fuzzy logic.
Responsibilities: - Implemented first derivative algorithms. - Implemented fuzzy logic in fuzzy control language. - Made a comparison between different algorithms to determine the best suited algorithm to detect edges.
Description: A project involving study and development of an algorithm to detect edges of grayscale images using first derivative algorithms and fuzzy logic.
Responsibilities: - Implemented first derivative algorithms. - Implemented fuzzy logic in fuzzy control language. - Made a comparison between different algorithms to determine the best suited algorithm to detect edges.
Payroll System
Environment: Java
Description: A simple java applet application to keep record of employee details and helps us to calculate house rent allowance (hra), dearance allowance (da).
Responsibilities: - Created an applet to accept the details - Implemented the project using GUI
Description: A simple java applet application to keep record of employee details and helps us to calculate house rent allowance (hra), dearance allowance (da).
Responsibilities: - Created an applet to accept the details - Implemented the project using GUI
Publications
Team Learning from Human Demonstration with Coordination Confidence Environment
The Knowledge Engineering Review, Cambridge University Press (pending publication)