EDUCATION
MS Info Systems, CORNELL UNIVERSITY
August 2019 - June 2021
-
MSc in Info Science and Info Systems, concentration in CS
-
Cornell Tech Merit Scholarship Award
-
Computer Vision researcbon 3D pose estimation using optical flow
-
NLP research detecting political bias using deep learning
-
Product Design Consultant For Estee Lauder Companies Online
BTech Comp Sci,
MANIPAL INSTITUTE
OF TECHNOLOGY
August 2014 - May 2018
-
BTech in Computer Science and Engineering
-
Graduated Top 10% of the class
-
NLP research on Morphological analysis of Kannada, an Indian agglutinative language, Paper Presented at ICCET 2018
EXPERIENCE
AI Product Manager,
AUDIBLE-AMAZON
April 2022 - Present
-
Drove 3.9x increase in annual revenue generated by the Content based AI team from $10M to $39M
-
building foundational services that support fast and efficient AI model development, data collection tooling, and internal productivity tools
-
Integrated Content AI assets into customer-facing features for content organization, personalization, recommendations, and search
-
Selected as one of two PMs for Gen AI taskforce to define Audible's strategy, prioritize foundational components, identify risks, and spearhead customer-facing AI features, reporting back to Audible's Cheif Product Officer and Chief Technology Officer
Software Engineering,
AUDIBLE-AMAZON
August 2021 - April 2022
-
Joined Audible Voice Experiences team responsible for playback of audiobooks through voice controlled smart devices(eg. Alexa)
-
Implemented an architectural switch from a ranking algorithm to a candidate proposer which given a customer utterance,
understands what is needed and which service should be invoked, improving the efficiency from 96.7% to 99.9%. -
Implemented an NLP model using Named Entity Recognition, LDA, Relation Extraction to auto generate the quiz questions
Software Engineering,
MICROSOFT
June 2018 - July 2019
-
Engineered pipelines that efficiently and compliantly extract Office 365 data on the Microsoft Substrate AI cloud platform
-
Extracted data was made available to partner ML teams as structured schematized data available for consumption
Research Intern,
MT-NLP Lab, IIIT-Hyderabad
January 2018 - May 2018
-
POS tagging of Konkani (a data-scarce language) with 77% accuracy using semi-supervised ML like active learning and self-training
-
Error Correction of Hindi POS tagged data with 94.02% accuracy
-
Model was built using a combination of support vector machines (SVM), long short term memory (LSTM) networks, hidden markov models, and conditional random fields to detect errors in Hindi POS data [Paper]
PAPERS, PUBLICATIONS & ACHIEVEMNTS
-
One of only 900 attendees globally, and the only Audible representative at OpenAI Dev Day
-
Automated Error Correction & Validation for POS Tagging of Hindi, ACL 2018, Paper presented at PACLIC 32, Hong Kong
-
Kannada morpheme segmentation using machine learning, IJET 2018, Paper Presented at ICCET 2018, Chennai, India
-
Patent Filed through Microsoft for a backend feature that uses NLP to improve the management of online communications
-
Served on the Judging Panel for National Level Smart India Hackathon 2020 hosted by Manipal Institute of Technology, India