Self Help App - PUSH-D #
Tech Stack: Java, Spring, React
Timeline: Feb 2022 - May 2022
- The National Tele-Mental Health Programme was announced in the Union Budget 2022-23 with the objective “to better the access to quality mental health counselling and care services”, by Nirmala Sitharaman, Minister of Finance of India. NIMHANS was identified as the nodal centre and International Institute of Information Technology-Bangalore (IIITB) was identified to provide technology support. This program is also being referred to as “Tele-Manas”.
- Developed a self-help platform where patients suffering from depression can seek Professional Psychiatric Consultation or follow the guided material present in the app.
- The webapp was composed of material that was provided by psychiatrists working in NIMHANS.
- A complete one stop solution for personalised management of Patients, Doctors, Specialist and other psychiatric articles for any consultation with the patient.
Railway Ticket Booking System #
Tech Stack: C, System Calls
Timeline: September 2021
- Implemented TCP Client Server Model Architecture for communication.
- Implemented File Management, Locking, Process Management, IPC mechanism to implement core functionalities.
- Socket programming to provide communication between server and clients.
- Optimized the system for multi-user access, ensuring data consistency through file locking mechanisms like fcntl().
- Admins can add/remove/modify trains, Agents can book tickets and Customers can update/cancel their bookings made by the agent for their account.
- The software supports Multi-tenancy and concurrency was maintained by file locking. All backend database storage is done in files and control is maintained by record locking.
Hydrocarbon Exploration using Seismic Imaging #
Tech Stack: Python, Flask, Keras Timeline: January 2019 - April 2019
- Designed and implemented a deep learning model based on U-Net architecture for seismic imaging to predict hydrocarbon deposits.
- Utilized Keras to develop a robust convolutional neural network (CNN) model tailored for 2D seismic image segmentation and analysis.
- Integrated a Flask-based web application to allow users to upload seismic images, process them in real-time, and generate predictive results.
- Processed and augmented large datasets of seismic images, optimizing the model for improved accuracy and detection of subsurface geological features.
- Deployed the model for binary classification of hydrocarbon presence, achieving 94.7% accuracy on validation data.