Team Lead for making model updates and rolling out new features in Gemini App Advanced subscription offering. Involves training numerous release candidates, and working with cross functional teams to ensure improvement of user experiences with every model update.

- TL for Google Gemini Advanced mode model updates- Making Google Gemini use Tools

Working in the Google Assistant Natural Language Understanding team.

Devising models for script learning in Natural Language Processing

Worked as a Teaching Assistant for the course "Symbolic Programming" that involves programming assignments in Clojure, a Lisp dialect.

Worked as a Teaching Assistant for the 'Neural Networks' undergraduate course at UT Austin which was taken by about 90 students. Conducted lab sessions and designed assignments to help the students familiarize with the implementation of different types of neural networks using libraries like PyTorch.

Developed a correlation service to analyze the impact of physical network faults on Google cloud services. Reduced the overall cloud outage detection time by about 40%.

Worked as a research engineer on a wide range of research problems with business use cases concerned with Adobe's Digital Marketing Cloud and Document Cloud. The problems related to the areas of Data Mining, Machine Learning and Natural Language Processing. Over the course of 2 years, I published 3 peer-reviewed conference papers , filed 5 patents with the USPTO and mentored 14 undergraduate Summer interns in various projects.The article attached gives the description related to one of my projects that was publicly showcased at Adobe's annual Digital Marketing Summit in 2017. It can also be accessed at Youtube using the link below.

Developed utility mobile applications that eased the life of students on the IITR campus. I also led the efforts in promoting the culture of mobile development in the campus by organizing open lectures and competitions as well mentoring junior members of the group.

Devised a novel solution to the problem of improving the on boarding experience of a new software product for a user using machine learning techniques to identify expert user patterns and using them to recommend to features to novice users.

Developed a low-cost Data Acquisition System (DAQ) based on BeagleBone Black for college students. We built a mobile application backed by a Node.js server for this purpose. The user could specify the configurations of the sensors interfaced on the board and observe the readings as well in a user-friendly manner.