Conducted research in developing scalable, differentiable algorithms for large-scale computer vision and discrete optimization tasks.

Worked in R&D department to solve imaging problems for ZEISS business units:1. Led project for 3D reconstruction. Delivered up to 3x improvement in accuracy by uncertainty aware fusion of depth maps from CNNs. Performed successful transfer of technology to multinational team of 3 developers.2. Led project for calibration of surgical robots. Gathered system specifications from robotics engineers and delivered an inexpensive approach. 3. Assisted in defect detection pipeline development for manufacturing processes using CNN. Resulting method alleviated need for laborious human annotation while also producing better results.4. Initiated studies for similarity search on whole slide images by self-supervised learning.

1. Devised cutting edge algorithms for detecting a variety of features in 3D seismic volume, and lead the team for implementation in the end-product.2. Provided consultancy in realizing technology roadmap.

My primary role was to perform research in image analysis algorithms for interpreting 3D seismic data. My main achievements were:1. Pioneered real-time image analysis software increasing throughput by 10x by parallel computing. Started as a sole developer, and later led a team of 5 team members to scale the solution. 2. Devised algorithms for tracking features in 3d volumes reducing human effort drastically. 3. Published algorithms for geophysical image analysis in international venues. 4. Led recruitment and training of 6 team members focusing on algorithms, image analysis and programming.