Jul 2021 – Jun 2023 · 1 yr 11 mos · Bengaluru, Karnataka, India

• I am responsible for productionizing machine learning models to ensure their scalability, real-time performance, and optimization.
• I successfully implemented solutions such as shoppable looks, vertical widgets reordering, beauty product care personalization and more

Feb 2021 – Jun 2021 · 4 mos · Bengaluru, Karnataka, India

• Worked on Automated Report key features extraction tool using pre trained Transformer model(Bert) with accuracy of 93% on very low dataset.
• The tool is currently used by GDSD team post successful training/ testing of the tool on real life weekly emails
• Worked on the Analytics of the Rakuten Point Reward System

Nov 2020 – Feb 2021 · 3 mos

• Researched a novel approach on project Sequential Pattern mining to extract high utility patterns from e-commerce user’s behaviour using Reinforcement Learning.
• Built a complete RL Environment to apply different RL algorithms making all the process automated and beating all existing methods.
• Filed a Patent for the given approach and writing a research paper for the top conference.

Sep 2020 – Feb 2021 · 5 mos · New York, United States

• Working on personalized AI fitness trainer for Yoga and various exercise pose correction.
• Building Deep learning models for detection for different correct poses.
• Researching on different state of the art pose estimation methods.
• Building algorithms to count reps of all types of excercise in edge devices.

May 2020 – Jun 2020 · 1 mo · Delhi, India

• Worked on different Reinforcement Learning Techniques to apply on image processing. Developed a pixel-wise agent, in which each pixel takes suitable action for improving the quality of the image.

Nov 2019 – Jul 2020 · 8 mos · New Delhi Area, India

• Collaborated to work on one of my ambitious idea Automatic Dubbing in which we built a system to dub English video to India Regional Language in real human voice
• Worked on various Text to Speech models like Tacotron and Flowtron
• Improved the accuracy of the Speech Recognition model using various different open-source tools

Sep 2019 – Jan 2020 · 4 mos · New Delhi Area, India

Dec 2018 – Sep 2019 · 9 mos

• Automatic License Plate Recognition - Trained Models like YOLO(Object Detection Algorithm) from Scratch.
• Speed Monitoring of vehicles - Autonomously developed a system to determine the speed of vehicles with 85% accuracy by using a single camera without calibration.