1. Dynamic SLAM with Deep Object Detection and Tracking for Moving Objects● Spearheading integration of deep learning models with Dynamic SLAM for enhanced detection in variable environments.● Implementing advanced tracking algorithms for consistent motion prediction of dynamic entities.2. Evaluating Alternative Pre-processing Pipelines for Object Detection● Leading evaluation of diverse pre-processing pipelines to optimize object detection performance..● Analyzing effectiveness and efficiency across multiple detection scenarios for robust real-time applications.

Serving as a teaching assistant for two courses, CS 4550 and CS 5610, providing academic support to students and assisting instructor with course logistics.Providing support for a class of 320 students by addressing queries and grading assignments for students as part of teaching assistant responsibilities.

● Data-Driven Decisions: Engaged with diverse retail data, such as sales, inventory, and customer behavior, to amplify operational efficiency and uncover business prospects.● Geospatial Expansion Analysis: Unveiled strategic expansion locales through geospatial clustering (DBSCAN, K-means) and regression models, driving a 25% business growth.● Stock Forecasting: Deployed ARIMA and LSTM networks for sharp stock predictions, reaching 90% accuracy and fine-tuning inventory management.● Automated Quality Control with CNNs: Improved wood quality control by using cutting-edge CNN architectures like VGG, ResNet, and EfficientNet. Used YOLO to locate defects in hardwood boards in real time, ensuring efficient processing and higher product standards.● Dimensionality & Anomaly Analysis: Integrated Autoencoders for compressing voluminous wooden plank data, enabling efficient anomaly detection and ensuring early identification of nuanced defects.● Performance Monitoring & Feedback: Leveraged metrics like MSE and MAPE to uphold model accuracy. Initiated feedback loops through periodic manual checks to iteratively refine model predictions.● Cloud & Infrastructure: Adopted AWS for robust and scalable model deployment, streamlining integration with business operations.● Collaborative Engineering: Promoted interdisciplinary collaborations, steering brainstorming sessions and aligning machine learning endeavors with overarching business goals.

● Collaborated with a team of 10 developers to enhance, modify and streamline Yardi's in-house product, VendorCafé, resulting in a 30% increase in user adoption and a 60% increase in customer satisfaction● Led the development of a VendorCafé feature using Java, Struts, MongoDB, ExpressJS, ReactJS, and NodeJS, enabling vendors to self-manage cost change requests, thus optimizing communication and approval workflows through automated updates and systematic record-keeping.● Built regression models to predict the impact of cost change requests on key project parameters like completion dates, resource requirements, and overall project costs