Experience

• 

  Humonics global

  Jun 2019 - Jan 2020

  Data science intern

  • Worked on a project called Automating Car Insurance Claims using Deep Learning• Built different image classifiers using fastai and PyTorch like car color, LHS-RHS, damage classifier, car profile, and car model.• Worked on Data Augmentation, preprocessing pipelines for training models and built Object Detection models using RetinaNet and YOLOV3.• Built a lot of semantic segmentation models like U-Net, FPN, PSPNet, DeepLabV3, and finetuned Mask-RCNN for car parts and damage detection.• Worked on a project called Document Understanding AI which involved building an OCR model using Deep Learning, creating pipelines and models for extracting useful information from documents.• Worked on reducing inference speeds for models at production using TorchScript with C++ and JIT. Show less

• 

  Sparrosense

  Jun 2020 - Aug 2020

  Research intern

  • Worked on Unsupervised Representation Learning for steel pouring classification in the steel manufacturing process.• Improved accuracy of classifiers from 85-93% and also improved robustness of the model to different scenarios in the industry.

• 

  Retail pulse

  Aug 2020 - now

  1. Worked on creating value from raw Kirana stores data.2. Worked with computer vision algorithms to get insights from data by classifying and detecting products in an image.

  • Machine Learning Engineer

    Aug 2021 - now

  • Research Intern

    Aug 2020 - Aug 2021
• 

  Kirana club

  Dec 2021 - now

  Machine learning