Agriculture consistently stands among the top three professions prone to disabling injuries and hazards. To assess the risk associated with maintaining odd postures for long period of time ergonomic assessment is used. Ergonomic design is used to enhance workplace productivity, workers’ comfort, and minimize injuries and hazards. Various musculoskeletal disorders can occur due to continuous use of odd postures, repeated activity and load carrying, torsional stresses. In this paper, we employ state-of-the-art deep learning model (YOLO-V11, YOLO-V8, YOLO-NAS) for keypoints detection. Based on the detected keypoints, angles between various body parts, such as the shoulder, elbow, and wrist, are calculated. These angles are then used to compute RULA (Rapid Upper Limb Assessment) and REBA (Rapid Entire Body Assessment) scores, which help evaluate the risk levels associated with the worker's posture. The RULA (Rapid Upper Limb Assessment) and REBA (Rapid Entire Body Assessment) scores derived from the posture were utilized for analyzing the risk. The results show that YOLO-V11 gives best performance with Average Precision of 95.6% and Average Recall of 95.7%. We also experimented with YOLO-V8 and YOLO-NAS, YOLO-V8 also gives impressive results.