Amit Kumar Patel

A Next.js application that allows users to upload files and chat about the contents of the files. This application is highly scalable across several screen sizes. The UI is lightning-fast and highly intuitive. A minimalistic, clean design is used for a feature-rich application. The file preview is implemented, and the PDF is viewed using the open-source PDF.js library, which is customized to support multiple string searches simultaneously. Moreover, the application handles failures and enforces error boundaries to render relevant error pages.

A React application for compliance management. It was intended to manage the legal compliance of companies and used to keep track of the transactions of compliance officers through the data from Plaid. It consisted of various restrictions and settings that would control the transactions and notify of violations. It allowed uploading marketing materials that would pass through several questionnaires for approval. It also used AI to highlight important chunks in the marketing materials. The system allowed creating certifications for various purposes based on the existing or new templates that could be created within the application. The certificate could be circulated and validated after being signed.

The application could be tailored for various organizations using the same core compliance engine. It was being tailored for over five organizations. Python was used for developing the REST API and had strict role-based access control. The managers could also impersonate an employee and perform actions on their behalf. It was a fairly large project with several modules and numerous features.

A React and PHP application aimed at keeping track of company projects, their resources, sprints, tasks in the sprints, time tracking on the tasks, and project reporting.

The front end was developed in React, and the back end was built using PHP and the Laravel framework. The application can register all projects and assign resources to the projects. Each project can have sprints, and each sprint can have tasks and epics. Tasks can be set to one or more resources and can log their time. The tasks assigned to resources can be added to their timesheet.

Each project can generate weekly reports and track the project status, health, and progress. The managers were able to import a large number of tasks from Excel files and auto-create tasks and sprints from them based on a pre-populated template for each project dynamically.

The application is being used by the company and is also a product to be marketed to other companies. It is on a subdomain and only accessible via VPN.

A Node.js application for contactless pickup service during the COVID-19 pandemic. As a full-stack developer, I looked after the implementation of the retail order lifecycle and the back office interface. I also developed the order interface that the end customers would view. MySQL was used as the database, Sequelize as the ORM, and Express.js for the server implementation. Embedded JavaScript Templating (EJS) was used for the order interface.

Whenever an order is generated, customers can track it via the order interface, and the order is processed via the back office. After the order is ready, the customers schedule an arrival time based on the available time slot and the slot capacity. They are provided with the route to the location via Google Maps and the estimated arrival and waiting time shown on the order interface.

The back office shows all the orders with relevant filters and notifications. It includes insights, several business indexes based on locations, and overall indices.

The stores are distributed across 10+ time zones and have time-critical requirements. More than 1,000 stores still use this application without any bugs to date.

A React Native app with Python (Flask) API to monitor the daily batch run of clients. The app provided real-time monitoring of the ETL batch processes running on the Snowflake database and notified users at various stages of the batch process, including any failure with details. This app provided analytical visualization of the runtimes and was used to identify bottlenecks in the batch process. It provided a graphical way to analyze data over chosen time duration with an ability to interact with the data points and dig down into the details of that point.

The back end was developed in Python (Flask) with JWT authentication. The Python back-end would monitor the Snowflake database, pulling data and converting it to Pandas DataFames and comparing it with the local copy, and triggering notifications via Firebase for any useful changes. The back end would only monitor the ETL batch process during the predefined batch run time controlled by crontab settings. This crontab setting would start and close the monitoring of various clients based on their batch runtime configuration and time zone.

This application helped the clients improve their ETL scripts and pinpoint the scripts that were taking the longest time.

An Android application developed to create a community of consumers and businesses brought together to create a better life experience. Based on the purchases using the account added to this application, one could rate the local businesses. Consumers and business owners could communicate directly via the app. Consumers could earn rewards for their ratings, while business owners could buy memberships to get their business promoted and highly visible.

In-App Purchase was used to provide various membership plans to the business owners. Google Maps API was used to locate the businesses. The app was released on the Google Play store and was particularly intended to be used in several states of the United States.

A Python (Flask) campus project for issuing and verifying digital certificates over the Ethereum blockchain.

Any institution could issue digital certificates to students or employees in a JSON file format. Each file would be a part of a batch of certificates issued at some time. The hashing of each certificate in the batch would generate a Merkle tree whose root would be stored on the Ethereum blockchain. The certificate would have the hash of the adjacent certificates on the Merkle tree, used to generate the Merkle root. Transactions are retrieved by using the transaction ID on the certificate. The Merkle root could be verified using the information of the adjacent certificate hashes on the certificate and its hash, which prevented any tampering with the digital certificate and guaranteed that this certificate was indeed a part of the batch issued on the blockchain network using the specified public key.

A centralized body deployed a smart contract to verify that the public key belonged to the claimed institution when the certificate was issued. The smart contract details were included in the certificate for verification.