Contributed to the Clinic Operations portion of the Whiterabbit.ai → RadNet (DeepHealth.ai) acquisition by migrating CRM cloud services from a Whiterabbit.ai-managed AWS environment to a standalone AWS infrastructure intended for use by the acquiring party. Led the replacement of SignNow with Adobe Sign to align with DeepHealth.ai’s integration standards. Improved system reliability by developing extensive unit tests, raising code coverage from 39% to 87% and meeting acquisition-mandated code quality requirements.

Customer Relationship Management (CRM) tool

• Developed both frontend & backend of CRM tool to support whiterabbit.ai's screening mammography compliance program (using Django framework, google-auth, celery workers, Redis cache, Postgres database) involving a campaign of 9 stages and a UI to display next three customers to call

• Upgraded CRM by enabling real time locking and notifications (using WebSocket, powered by ASGI server Uvicorn) to implement a shared view of tasks across agents and a UI to handle incoming messages in real time

• Build features to manage agent activity on CRM and implemented security features involving a cache queuing mechanism to restrict agent to a limited pools of tasks at a time, different agent statuses & inactivity auto-idle

• Supported development of calling module in CRM using third partly application “twillio” and “bandwidth” to perform outbound calls directly from CRM; extended the feature by adding functionality to receive incoming calls

• Supported development of the data ingestion module of CRM to efficiently ingest data from different RIS systems using provided SQL Stored Procedures, direct database queries, HL7 messages and spreadsheets

Mammogram Booking Link

• Implemented booking link (to be sent in text messages as a part of campaign) to enable customers place a mammogram appointment request for a real appointment slot, in eRAD RIS, at a particular clinic

• Developed algorithms to securely locking the selected slot for 30 minutes strategically using multiple eRAD APIs

• Demonstrated the UI and locking mechanism to the client and launched the product enabling 90K+ clinic customers to book a mammogram appointment directly into the eRAD (and converting existing pre-orders to bookings, if applicable)

Curofy iOS App (an app that enables doctors to coordinate with other doctors on patient cases)

• Worked as a developer for iOS App of Curofy, a platform on which doctors can discuss cases, read latest medical news, search for jobs, and have access to latest medical articles and updated medical guidelines

• Added features like sponsored carousals, library (articles, guidelines and MCQs), leader-board, image-editor and modified almost all old features including doctor’s profile, discussion feed, and notifications

• Programmed in Objective C; used Cocoa Pods libraries like Masonry for design; worked on caching and firebase

Curofy Backend

• Developed backend APIs using Django framework for library feature (comprising of MCQs, medical journals, medical research papers and articles) on app incorporating pagination, sorting and efficient search of text (across all the articles and journals)

• Created a parser to crawl NCBI website to fetch medical articles, thereby eliminating the need to upload articles manually and improving the frequency of article input from 6-7 per day to 40-50 per day

• Added various features on Curofy admin panel to provide technical help to content managers; automated the tasks of tagging medical articles, news and medical cases (posted by a customer).

• Created a chrome extension to fetch medical content from a single-page application and export to app directly, further reducing the work of the medical content team

• Automated task of fetching latest medical news on an hourly basis from 4 online sources and formatting it to display on Curofy App with proper medical tags.

Medical Image Classifier

• Used TensorFlow and data of cases posted, on Curofy, to classify any image amongst categories including X-Ray, ECG, Medical Poster, CT Scan, MRI and non-Medical

LinkedIn Aspire

• Worked as a Backend developer to analyse, rate and provide suggestions to improve LinkedIn public profiles

• Extended the task of building a chrome extension to crawl and fetch LinkedIn public profiles

• Developed algorithms leveraging techniques of NLP and Artificial Intelligence to extract skills from LinkedIn profiles

VMOCK Scrubber

• Created an offline JAVA application, with various security features aimed at removing personal information from resumes; to enable students of business schools share their resume with VMOCK without sharing any personal information online. Launched the product for students and alumnus across top 4 global business schools in the US