Multi Object Tracking (MOT) plays a crucial role in various software applications, including autonomous driving, security camera systems, and athletic performance analysis. Despite significant progress, problems like occlusion, irregular object movement, and maintaining real-time performance persist. Our approach integrates Detection Transformer (DETR) for robust object detection with an adaptive Kalman filter for accurate adaptive motion prediction. To deal with the complexities of overlapping detections and reduce or completely remove identity switches, we propose a custom Intersection-over-Union (IoU) loss function specifically tailored for challenging tracking scenarios. With the strengths of the Detection Transformer (DETR), our method achieves high-accuracy object detection, while Kalman filters ensure consistent tracking under diverse dynamics. Our tracking framework is trained upon and evaluated on datasets such as MOT16, MOT15, MOT20, and MOT17 which bring forth diverse and challenging scenarios for Multi-Object Tracking, diverse evaluations of the MOT17 benchmark demonstrates the efficacy of our system, this achieves a competitive Multiple Object Tracking Accuracy (MOTA) score of 62.7, achieving scores better than baseline models such as TrackFormer and Tracktor++ for on-demand apps. This result highlights the scalability and reliability of our framework in handling complex and diverse tracking environments. Our contributions provide a solid base for the upcoming MOT methodologies, in inference for both academic research and real-world deployments and hence the title, Multi-Object Tracking Using Detection Transformer and Adaptive Kalman Filtering.