-
Experience
Associate Trainee (Software), Apps Associates, Hyderabad, India Aug 2017 – June 2018 (10 months)
• Developed dashboard for internal team analytics
• Migrating CI/CD pipeline for projects and Writing Integration tests for services in projects.
• Developed internal web application which involves java, Spring MVC and hibernate.
• Added functionalities using Java and J2EE technologies and developed Rest API’s using Spring and Nodejs
• Designing database schema and writing efficient SQL queries. Front-end engineering using Java script and libraries.
• Participated in developing code for adding new features and removing bugs from different MVC projects.
• Worked of AWS stack (including EC2, Route53, S3, RDS, Dynamo DB, IAM, SQS, SNS, LAMBDA).
-
Projects
The current standard for prehospital communication with Emergency Department (ED) personnel is to use a system called IMIST-AMBO. This Capstone project will use recorded pre-hospital phone call record files saved at Phoenix Children’s Hospital Trauma and
1) convert the speech to text ,
2) classify the text into the IMIST-AMBO categories.
This application will minimize gaps in critical pre-hospital information needed to save lives from traumatic injury. The application may also lower medical errors and costs considerably. Users can use this application to identify all the IMIST-AMBO categories which is useful for them to treat the patient, if some part of IMIST-AMBO is missing then the user can immediately ask the responder about the missing parts which are essential.
To achieve this, we have created a web-application using React on Frontend, Python’s Flask on Backend and MongoDB as the Database. We have used AWS S3 Buckets to store the phone call recordings. To convert these audio recordings to text we have used AWS Transcribe Tool[1]. For the second part, which is classifying the text, from the speech-to-text conversion, into IMIST-AMBO categories, we have used python’s NLTK library[3] and our own rule-based Machine Learning algorithms to achieve effective categorization. We have deployed the web app in production using AWS Cloud Resources like VPC, Security Groups, EC2 and Elastic IPs.