Experience | Neha Boloor

ML Engineer

Productionizing Generative Driving Scenarios to enhance testing and validation of the robot's prediction and planner stack by augmenting & fuzzing vehicle logs for diverse scenarios.

Designed and developed a frontier-based autonomous exploration algorithm, transforming the camera-powered robot vacuum from semi-autonomous to fully autonomous exploration.
Optimized the algorithm to achieve a remarkable 3x increase in speed, enabling exploration of an average 2000 sqft household in ~15 minutes.
Crafted an end-to-end simulation test framework with Unreal Engine (UE) and Grafana for iterative improvements through user feedback, A/B testing, and hyperparameter optimization.

Improved depth prediction along the image edges and for low-light conditions through data augmentation.

Designed and developed key marketing segmentation, data integration, and performance enhancement features for Oracle Marketing Cloud's SaaS products.
Collaborated and set up an automated testing framework to enable the product's Quality Assurance Team.

Explored the suitability of Faster R-CNN for the purpose of CBIR under the guidance of Prof. Anurag Mittal.
Achieved 4.5% increase in mAP with data augmentation, triplet network and self learning classifier.