Mayank P.

Senior Deep Learning Engineer @AstraZeneca

Using graph convolutional neural networks to learn deep representations of gigapixel histopathology images- Using deep convolutional and attention based neural networks to learn deep representations of CT volumes. Using attention based models to improve performance for subsequent downstream tasks- Writing data quality control tools for pre-processing medical image data and providing it efficiently for neural network training. Reduced disk and memory usage required by around 25%, allowing increase in network training bandwidth- Communication with clinical stakeholders, and using their feedback to direct improvements in future research and engineering tasks- Conducting technical outreach by preparing and submitting results to academic journals and conferences, and delivering presentations about technical topics- Onboarding and mentoring new joiners, junior engineers, and interns.