Hiresh Gupta

Hiresh Gupta

CV Grad Student at CMU | ML@Apple | Adobe | BITS-Pilani (CS)

Apple, Cupertino

Machine Learning Intern (May’22 - Aug’22)

  • Worked with the 3D Vision team at Apple Maps to detect road features like lane markings, sidewalks, bike paths, etc.
  • Incorporated satellite imagery as a new data source to segment road features and update road network information.

Adobe Systems, India

Machine Learning Engineer-II (Jul’18 - Aug’21)

  • Designed multiple network architectures to digitize print forms by extracting the hierarchical document structure.
  • Proposed a high-resolution segmentation method to extract the form structure to up to four levels of granularity.
  • Developed a service to directly transform print forms to responsive HTMLs for better viewing experience across devices.
  • Utilised conditional GANs to generate synthetic training data corresponding to certain complex layout patterns.
  • Scaled deep-learning workloads to TPU pods to achieve a 15x speedup in throughput and a 10x speedup in training time.
  • Designed multi-modal solutions to leverage spatial and textual information in documents to further improve the form structure extractor accuracy by 5%; deployed model to production.
  • Focused on customer-oriented research to make print documents re-flowable by deploying several deep learning models in AI Powered Automated Forms Conversion Service.

Media and Data Science Research Lab, Adobe India

Computer Vision Research Intern (Jan’18 - Jul’18)

  • Worked on a deep learning based Visual Product Search which segments different products in a wild image and performs search in a large catalogue.
  • Proposed a novel grid based training of siamese networks, allowing it to observe multiple positive and negative image instances simultaneously. The research was awarded the Best Paper Award at CVPR 2019 Workshop on Fashion and Subjective Search.

Samsung Research Institute, Noida

Software Development Intern (May’17 - Jul’17)

  • Contributed for development of a new Bluetooth Low Energy Profile on Air Toxicity, to create a standard platform for all air quality sensing devices.

Indian Institute of Remote Sensing, ISRO

Research Intern (May’16 - Jul’16)

  • Worked on the Development of Data Cube for Uttrakhand Region in India, to store and retrieve large Earth Observation datasets in an efficient manner.
  • Conducted extensive research on different compression techniques and data storage formats for efficient query, retrieval and storage of data.