Python OpenCV YOLO TensorFlow Keras Numpy Docker
We led the architecture and implementation of a camera-based, machine learning-driven vehicle alignment system designed to replace legacy motion sensors. The system utilized advanced computer vision techniques and real-time object detection models to ensure precise, automated vehicle positioning within pallet boundaries. This initiative aimed to enhance operational efficiency, safety, and reliability in automated parking facilities.
We integrated high-resolution camera streams with machine learning frameworks for occupant and boundary detection, alongside image processing pipelines that processed frames at scale. Robust alert mechanisms were implemented to provide immediate, actionable insights into alignment issues or potential obstructions. This approach minimized manual intervention, enabling smooth, continuous operation and improved user satisfaction.
The solution leveraged Python, OpenCV, YOLO-based object detection, and scalable backend services, delivering a cost-effective, extensible platform tailored for the evolving needs of automated parking environments. This project exemplifies my ability to lead complex, ML-focused initiatives that drive significant improvements in system performance and stakeholder experience.
"We were thoroughly impressed with the innovative approach and execution of the camera-based, machine learning-driven vehicle alignment system. The integration of advanced computer vision techniques and real-time object detection models delivered precise, automated vehicle positioning, greatly enhancing the efficiency and safety of our automated parking facilities. The solution's scalable architecture, robust alert mechanisms, and cost-effectiveness exceeded our expectations, minimizing manual intervention while ensuring continuous operation. This project showcases exceptional technical leadership and a commitment to delivering impactful, high-performance solutions"
PHP AngularJS MySQL jQuery REST API ElasticSearch AJAX
We designed and developed a robust e-commerce platform for Executive Advertising, handling both frontend and backend development. The platform featured an advanced search module powered by ElasticSearch, user analytics for behavior tracking, and administrative tools for order management, email campaigns, and product display.
Key functionalities included a detailed cart and checkout process, a product review system, and automation tools for database backups and data imports using cron jobs. The platform was hosted on AWS, with optimized web and database servers to ensure scalability and performance.
The successful deployment improved user experience, increased sales, and streamlined business operations. This project showcased my ability to deliver comprehensive, scalable solutions that meet both user and business needs.
"The platform exceeded expectations in every way. Its advanced search capabilities, user analytics, and streamlined administrative tools greatly improved business operations. The optimized hosting and automation features delivered scalability and top-tier performance, enhancing user experience and driving sales. An impressive and impactful solution overall"
Rosetta Stone Language score API integration. Oauth2 API development for authorization in laravel. Advanced Vuejs implementation for microphone setting
"The integration of Score API, OAuth2 authorization and advanced Vue.js for microphone settings brought exceptional functionality to the Rosetta Stone language learning system. A seamless and innovative solution overall"
Comprehensive chrome extension for fetch data from API & then manipulate admin panel data periodically. Redeem/Recharge account. Data exchange between tab & extension popup
"The chrome extension delivered seamless functionality, efficiently fetching API data, managing admin panel tasks, and enabling smooth account operations. The integration of data exchange between tabs and the popup was both intuitive and impactful—a highly effective solution!"
An online marketplace for condominium, single-family, and townhomes with interactive features search and compare properties, contact developers or brokers, and proceed through the purchase. Major components include an interactive and map-based searching and filtering module to find properties by location, occupation date, amenities, and more. Developed several modules including user login, roles, and permissions, dashboard, reporting, timeline features for prospects, realtors, agents, developers, and admins. Implemented content and media management dashboard, floor plans with 3D views. Implemented worksheet submission and management wizard
"PrismTecHub made an impressive and comprehensive online real estate marketplace! The blend of interactive features—like the map-based search and 3D floor plans—ensures a user-friendly experience, while the robust modules for user management, dashboards, and reporting showcase a well-thought-out and efficient system. By catering to a wide range of users such as prospects, realtors, agents, and developers. This is truly a powerful solution that merges functionality with innovation"
Wordpress Custom post type development including search feature. Data syncing facility with API
More detail: GIT Repo
"The WordPress plugin's custom post type development and efficient data syncing via API brought seamless integration and enhanced functionality. An impressive and highly effective solution"
Python Deep Learning Autoencoders Pandas Numpy SPRT algorithm
Compromised IoT devices are the main origin of botnet attacks. Recent DDoS and other botnet attacks have caused significant harm, and have been facilitated by the use of IoT-based botnets. Due to the limited human interaction required for IoT devices to function, it can be difficult to monitor all of their activities. Therefore, it is crucial to identify compromised devices within a network. My model uses Deep autoencoder to learn the regular network behavior of IoT devices in a network and continue monitoring the network. If a device is compromised, it can be detected as fast as 7 network packets with an accuracy up to 99.98%
"Impressive work! Your project addresses a critical issue in IoT security with a highly innovative solution. The use of a deep autoencoder to monitor network behavior and detect compromises so quickly, with such high accuracy, is remarkable. This approach holds great potential for mitigating botnet attacks and protecting networks effectively. Well done"
