This project involves developing a content management system for handling engine documents at Rolls-Royce. It includes a feature where users can interact with a chat window to request document references within a web application. The system utilizes the Llama2 model for document inference and exposes the model through FastAPI for integration with other services. Additionally, the model was optimized to enhance throughput and reduce computational resources, significantly decreasing response times from 30 seconds per request to just 2 seconds per request.

The project, developed for MAN Truck, centers around identifying competitive tenders across the EU. Its innovative design reduces the need for human resources by automating the process of checking deal conditions. Key functionalities include data crawling from public websites for tenders and enabling users to input search criteria such as keywords, location, and language. Additionally, users can utilize a chat window to request document references via the web application. The project employs Llama2 to filter and identify deals that align with MAN's specific conditions and exposes the model via FastAPI for integration with other services. The optimized model can efficiently process 2000 rows of data per minute.

A system designed to manage the content of product manuals, incorporating version control for documents and providing an offline mode through Progressive Web App (PWA) capabilities, allowing users to access information without an internet connection.

An online coaching platform offering a wide range of categories, such as Front-end, Back-end, DevOps, etc. It includes user management with access control lists (ACL) and the ability to manage actions for each user. The platform provides exercises for students and offers mentorship support. It has an intelligent agent using KNN to assign tasks to qualified mentors. Additionally, the system manages mentor income and allows mentors to create and oversee their own courses.

ToolScout is an intelligent tool selection platform designed for quick and application-specific tool selection. It includes features for selecting tools across various manufacturers and standards, calculating application data, providing reliable engineering tool information, and calculating cutting data. The platform also allows users to configure their working environment on the website, including personal machine tasks and tool favorites.

Developed a robust system for a car booking service, which includes a comprehensive database design using MongoDB. This project involved the creation of separate mobile applications for users and drivers, enabling real-time notifications for drivers regarding booking requests. It also integrated a payment gateway to support various payment methods.

This project involves managing the card box system within a casino using the Jarvis ecosystem. It includes tracking the location of card boxes in-house, developing an app to unlock card boxes using NFC and BLE, monitoring the lock/unlock status of boxes, analyzing the number of card boxes, and utilizing MQTT to subscribe to and publish data, providing alerts to managers.

This project involves developing a system to identify spelling and grammar issues using Natural Language Processing (NLP). It includes preparing a dataset for training the model, researching suitable network architectures, and fine-tuning the model to optimize performance.

The Jarvis Eco-System Visitor Management project is designed to oversee visitors, manage parking spaces, and monitor vehicle occupancies. It utilizes camera technology to scan vehicle IDs, tracks the number of vehicles in the parking area, and manages overnight parking. The system also organizes visitor access within different building zones.

This project is designed to monitor various environmental parameters such as the amount and pH of rain, water levels, and the temperature in drainage systems. It sets thresholds for these parameters, and if any value exceeds or falls below these thresholds, it alerts users, aiding in weather forecasting. Additionally, it analyzes data to predict potential floods or storms.

This project is designed for marketing within a mall environment. It allows visitors to explore promotions and stores categorized through an AR-powered indoor navigation system built in Unity. This system is integrated with native iOS functionality and utilizes beacons to accurately determine user locations indoors. An admin page is available to help mall owners track foot traffic over specific periods such as daily or weekly intervals. Owners can also manage beacon ranges and devices.

This project involves developing a system that assists nurses in monitoring elderly individuals. It provides features to track the health and location of elderly patients, manage room security and devices, and send emergency alerts using real-time data tracking with MQTT technology.

This project offers features for users to manage their tasks, book rooms, and communicate with each other. It uses Redux for global state management, validates user input, and provides warnings when tasks do not meet their timeline. The real-time chat is implemented using Socket.io, and there are alerts for unauthorized room access. The project also includes a mobile app that users can use to scan QR codes for booking validation, and it tracks user locations within rooms using BLE.

An ecommerce ecosystem project focused on designing and implementing a backend infrastructure using MongoDB and PostgreSQL databases. The project involves creating RESTful APIs with a microservice architecture to ensure high availability and performance. Additionally, it includes an automated agent to assign orders to stores and coordinate delivery drivers.

This project involves the development of a comprehensive website for a coffee shop, providing customers with detailed information on products, store locations, and the latest coffee lifestyle trends. It includes state management through Vuex, real-time cart synchronization with the backend, user behavior tracking via Google Analytics, database triggers using Firestore, and integration with various payment gateways.