1. Using Transformers, LLMs and Generative models to advance Sponsored Ads on Amazon.2. Working on NLP and CV solutions such as machine translation, layout generation and image generation to generate and translate advertising assets (text and images) to other languages, backgrounds and geographies. 3. Improving relevance and coverage of Sponsored Products search results on retail websites worldwide (eg: Amazon.com, Amazon.ca)

Working on the DARPA ASIST project with Prof. Katia Sycara in AART Lab.Built a multiagent gym environment to learn and study group cooperation strategies for mission critical tasks.https://www.darpa.mil/program/artificial-social-intelligence-for-successful-teamshttps://www.ri.cmu.edu/robotics-groups/advanced-agent-robotics-technology-lab/

Trained LLMs and Generative Adversarial Networks to generate creative banner layouts and layout variations given input elements such as text, images and shapes. Experimented with 3 public and 1 private dataset.Filed a US patent for the LayoutGPT variation that I developed.

Media and Data Science Research Lab, Digital Experience1. Fundamental and Applied Research in Machine Learning2. Applying data science techniques to digital marketing to improve customer engagement and retention.2. Led several pilot projects across 9 European countries to personalize email marketing campaigns for a $300 billion net worth company.3. Developed and shipped the algorithm to generate Personalization Insights Reports in Adobe Target.4. Led research projects in the fields of Reinforcement Learning and Explainable AI having applications in digital marketing.5. Delivered talks, paper presentations and lab sessions at Adobe Tech Summit (2019 and 2021), International Conference on Learning Representations 2019 (New Orleans, USA) and Quantum Techniques in Machine Learning 2019 (South Korea)6. Published papers, filed patents, interviewed and supervised new-hires.

Continuing work on the research problem that I worked on in the summer.Currently carrying out extensive literature survey for the field of interpretable AITrying to extend interpretable AI to the fields of regression and natural language.

Worked in the field of Interpretable AI, employing the use of shapley values, linear models and evolutionary algorithms to build if-then rules that give insights into black box machine learning model behaviour. Solved it for classification models.Made User Interface using Python Django. Published at ICLR Workshop 2019

Worked on comparing 4 swarm-based Q-learning algorithms for a grid world task.Implementation was done using MATLAB .Worked on text summarization techniques that used neural networks.Implementation was done using Python.Subsequently wrote a paper that has been attached.Has been accepted to the 18th International Conference on Computational Linguistics and Intelligent Text Processing