An app for managing online classes. Made with flutter and Flask+ SQL backend.

Hosted at:

Synchronize your sticky notes collection across all platforms, and save ideas on the go. A flutter projects built and released for windows, android, ios , and web as well.

Hosted at:

This is a file-sharing utility that enables cross-platform, secure and fast file sharing between two devices. Mainstream file sharing choices like whatsapp are built as client-server, which makes sense for instant messaging but not so if purely for file-sharing. The technology used is WebRTC for p2p connection and data transfer, react JS for website, and capacitor by Ionic for converting the same to an android app (cross-platform). For the signalling/handshake between two devices firebase acts as intermediate/signalling server. But once the two peers make a connection, firebase loses all session information and the file-share phase begins

Hosted at:

An app for a water purification and delivery company to manage orders digitally. (Enterprise resource planning)

Hosted at:

An Internship at HAL Mission Combat Systems Research and Design Center (MCSRDC) division. HAL excels in producing different aircraft models and manufacturing them for defense purposes, and as a part of this airborne model design process, the virtual 3D models are simulated in the HAL proprietary Simulation engine. However running this simulation manually by manipulating various parameters is a time and effort intensive process. I was part of an intern group working under the mentors responsible for this, and we were assigned the work to automate and thus reduce manual labour in this testing phase. The devised solution was to have a UI/ interface that has various choices for simulation environment and 3D models to be tested within that. Also templates of set of testcases to be executed for given airborne model could be individually created and saved. These templates can be simply loaded as and when needed and customize parameters for slight changes in requirements. This helps avoid redundancy as across consecutive tests a large portion of parameters stay the same. Particularly i was responsible for handling the API interface between our UI and the engine for hardware linking. In short, over a duration of one month, the task involved was to automate integration testing as a part of pipeline. Ultimately, this reduced manhours for testing new models by 17.8 percent.

Hosted at:

An LSTM model that does demand forecasting and suggests inventory decisions.

Hosted at:

A solution with the aim of reducing farmer suicides in India. In the contemporary system, farmers borrow money in the form of loans from moneylenders/banks. This leads to additional financial burden on farmers in case of crop loss due to natural or any other causes. Current suicide rates in India are a huge concern. We thought of an investment model platform instead of loans for farmers/land owners. Investors and land owners can sign up accordingly and different farmlands are listed similar to how stocks of different companies are listed in trading sites. The financial model was based largely on stock market model. Additionally, for investors tools like crop price prediction and to assist in their investing, we made a LSTM model that runs in the backend and extrapolates the fluctuating crop prices to next few days. The software was packaged as an android app as android phones currently has the most reach in rural areas in India. This app was built using React Native.

Hosted at:

A shopping assistant in the form of chrome extension + website for suggesting greener alternatives to the current product. The extension is compatible with common shopping sites like amazon, flipkart, shopclues, so on. The tool is responsible for getting the carbon footprint of the current product being viewed in any of the above shopping websites. Not just that, but fetch alternatives with lower carbon footprint if any of our partners provide suitable green alternative product. So from business perspective, it invloved partnering with businesses that aim for sustainablity and making the products appeal to customers. As an incentive for customers to try out these green product they get rewards in terms of personal contribution to carbon footprint saved.

Hosted at:

A deep learning model that is trained from datasets from the cancer imaging archive. Then make a UI for doctors to interact with model and utilise it to save their time and get classification and severity of tumour if present faster and hence act faster

Hosted at:

An ML model that uses anomaly detection for automating surveillance system and alerting forces.

Hosted at: