Sonali Sharma

Senior Engineer with hands-on experience at Bosch Global Software Technologies, focused on ADAS perception and performance optimization. Skilled in solving complex technical challenges and supporting Agile teams by informally stepping into Scrum Master responsibilities. Proficient in Python and C++, with a strong background in technical documentation, sprint facilitation, and cross-functional collaboration.

Image data augmentation

• Implemented augmentation techniques (e.g., flipping, rotation, scaling, noise injection) to enhance dataset diversity and improve the robustness and generalization of deep learning models for computer vision tasks

Labeling data for ML

• Manually annotated traffic scene images with pixel-level precision to create custom semantic segmentation datasets for object classification tasks. Labeled key traffic elements such as vehicles, pedestrians, road signs, lanes, and sidewalks to train deep learning models for scene understanding Focused on maintaining high annotation consistency and class separation without the use of automated tools, ensuring data quality for training segmentation architectures. The labeled data was used to improve model performance in tasks involving real-time object classification and urban environment perception

AEB and ACC for radars

• Developed and fine-tuned object tracking algorithms for automotive radar systems to support Advanced Driver Assistance Systems (ADAS) functions such as Automatic Emergency Braking (AEB) and Adaptive Cruise Control (ACC). Focused on detecting and tracking dynamic objects (e.g., vehicles, pedestrians) in real time, estimating relative velocity, range, and trajectory. Implemented filtering techniques (e.g., Kalman Filter) to maintain robust target tracking in noisy environments. Ensured accurate object state estimation to enable safe braking and distance control decisions under varying traffic scenarios.
• Owned the end-to-end delivery of the object tracking software as a Perception subsystem for customer integration. Collaborated with cross-functional teams to meet functional safety and performance requirements, conducted validation and verification, and ensured seamless integration with vehicle-level ADAS systems for production deployment.

KPI evaluation tool for tracking performance 

• Designed and developed a comprehensive KPI evaluation tool to monitor, analyze, and visualize key performance indicators (KPIs) for machine learning and software development projects. Real-time tracking and reporting through customizable dashboards to facilitate data-driven decision-making and continuous improvement. Collaborated with stakeholders to define relevant KPIs aligned with project goals and integrated the tool into existing workflows for seamless performance monitoring and evaluation.

• Professional Scrum Master 1 (Scrum.org Credential ID 1216547)