AI Engineer @ Mercer Mettl || Ex - AI Algorithm Engineer Intern @ViSenze | Deep Learning (Computer Vision) Intern - AI Research Intern @ ML Studies | Data Science Intern @MuseWearables | Junior ML Engineer @Omdena
3 Years of Experience
Meerut, Uttar Pradesh, India
+91 9**********
-
-
Not Available
Hi, Currently, I am working as an AI Engineer at Mercer | Mettl. Before that, I did an internship at ViSenze (Singapore Office - Remote), in the role of AI Algorithm Engineer Intern. I was engaged in the research and development of an intelligent visual search application and built projects like improving the fashion category tagging model using Active Learning. I did my engineering at the Indian Institute of Information Technology, Vadodara, with major in Computer Science & Engineering (2021) [CPI-8.65]. During my AI Research Internship at ML Studies, I worked on a research project to produce better 'Visual Explanations' of decisions made by CNN-based models. I also did a Deep Learning (Computer Vision) Internship at Machine Learning Studies where I worked on an object detection problem for a retail store, using Detectron2 and fine-tuned pre-trained Faster-CNN, Mask-RCNN, and Cascade-RCNN, for this. I explored the problem of domain shift and creating artificial datasets with diversity. I also performed a significant analysis for adding augmentations, views, and constraints. I used image processing and computer vision techniques like morphological processing, connected components, and grabcut to clean up data. I have also contributed as a Junior ML Engineer at Omdena, a global community, where we try to solve real-world problems with the help of AI. I have contributed to 3 Omdena projects. Highly motivated to use #AI4Good. In one challenge, the 'Fighting Illegal Dumping challenge, I collaborated with 50 other AI practitioners in partnership with the Trashout project. I led a subtask and developed models to identify if a shopping product is sustainable and environmentally friendly and to which recycling bin it belongs using deep learning and computer vision techniques. I have worked on several projects like Super-Resolution of an Image using GAN, Instance Segmentation on crack data using Mask R-CNN, Background Subtraction, Content-based Image Retrieval, analysis and times series based forecasting on Birmingham car parking dataset, car detection using YOLO, and implemented the GRAD-CAM with sanity checks (unofficial implementation of ‘Sanity Checks for Saliency Maps’ NIPS - 2018). Computer Vision Engineer Data Scientist AI Algorithm Engineer Deep Learning Engineer Machine Learning Engineer ( ML Engineer )
Mettl, SaaS/Cloud Product, Computer Software
ViSenze - AI for Visual Commerce
Machine Learning Studies
Mettl, ViSenze - AI for Visual Commerce, Machine Learning Studies, Machine Learning Studies, Omdena, MuseWearables
Job Title : Artificial Intelligence Engineer - Computer Vision
Company name : Mettl
Period : July 2021 - Present
Location : Gurugram, Haryana, India
Job Title : AI Algorithm Engineer Intern
Company name : ViSenze - AI for Visual Commerce
Period : January 2021 - June 2021
Summary : • Improved street-style fashion category tagging model for 2 critical bad cases.
• Improved fashion category tagging model specifically for a customer by enriching training data using Active Learning improved recall of the 'dress' category by 14%.
• Build a 98% accurate tagging model to filter sensitive bad cases from the detection model's predictions - foot/hand/eyes from shoe/gloves/eyewear detections.
Location : Singapore
Job Title : Artificial Intelligence Research Intern
Company name : Machine Learning Studies
Period : August 2020 - October 2020
Summary : - Worked on a technique for producing better ” Visual Explanations” for decisions from CNN-based models.
Job Title : Deep Learning (Computer Vision) Intern
Company name : Machine Learning Studies
Period : April 2020 - July 2020
Summary : - Worked on an object detection problem for a retail store, used Detectron2 for this, and fine-tune pre-trained Faster-CNN, Mask-RCNN, and Cascade-RCNN.
- Explored the problem of domain shift and creating artificial datasets with diversity to improve the recall score from 54% to 96%. Performed significant analysis for adding augmentations, views, and constraints, as well as utilizing image processing techniques like morphological processing, connected components, and grabcut for cleaning up data.
- Worked with a variety of state-of-the-art libraries like Detectron2 (PyTorch), Albumentations, and Apex.
Job Title : Jr. Machine Learning Engineer
Company name : Omdena
Period : November 2019 - August 2020
Summary : - Fighting Illegal Dumping with AI, (July - Aug 20)
Worked with 50 other AI collaborators in partnership with the Trashout project which aims to map all illegal dumps around the world and help citizens recycle more. Developed models to identify if a shopping product is sustainable and environmentally friendly and to which recycling bin it belongs.
- Applying AI to Understand the Sentiments, and Aspirations of Young People, (Jun - Jul 20)
Developed tools to analyze and understand the sentiments and aspirations of young people and performed a temporal analysis to understand how the sentiments have been changing over time, especially due to the COVID-19 Pandemic. The solutions will help Fondation Botnar to better understand how to support young people more effectively and catalyze appropriate initiatives.
- Renewable Energy Challenge in Nigeria, (Nov - Dec 19)
Worked with Nigeria-based NGO Renewable Energy Africa on building an interactive map for finding the most suitable spots for solar panel installations across Nigeria. The areas helped the policymakers in Nigeria to make better decisions.
Job Title : Data Science Intern
Company name : MuseWearables
Period : October 2019 - November 2019
Summary : - Fulfilled all data science duties for a high-end product based company.
- Created models that help to analyze the impact of an Instagram influencer on its audience.
- Worked on the age and gender prediction and language identification machine learning models.
Title : Convolutional Neural Networks
Period : July 2019 - Present
Summary : 9448V755UTX4, coursera.org, https://www.coursera.org/account/accomplishments/certificate/9448V755UTX4
Issuing Authority : Coursera
Title : Improving Deep Neural Networks : Hyperparameters tunning, Regularization and Optimization
Period : June 2019 - Present
Summary : JFF7LVMLHSVJ, coursera.org, https://www.coursera.org/account/accomplishments/certificate/JFF7LVMLHSVJ
Issuing Authority : Coursera
Title : Structuring Machine Learning Projects
Period : June 2019 - Present
Summary : PXA74AXFWXFZ, coursera.org, https://www.coursera.org/account/accomplishments/certificate/PXA74AXFWXFZ
Issuing Authority : Coursera
Title : Neural Networks and Deep Learning
Period : May 2019 - Present
Summary : FUL9EJ8XACYP, coursera.org, https://www.coursera.org/account/accomplishments/certificate/FUL9EJ8XACYP
Issuing Authority : Coursera
Title : Digital Image Processing (DIP)
Period : March 2019 - Present
Summary : UC-3APYVDK9, udemy.com, https://www.udemy.com/certificate/UC-3APYVDK9/
Issuing Authority : Udemy
Show More