Search improvement: Working with Product and Search AI teams to improve eBay core search efficiency. Data Science lead for the Germany market product initiatives

Kidas | Nov 2022 - Jun 2023

Machine Learning Engineer

● Core NLP: Conceived, developed and deployed several BERT based classification models, which were integrated into the company’s core product. The developed models identify sensitive topics from in-game conversations of children including personally identifiable information, hate speech, sexual content, bullying. These reduced manual overhead by 60% and improved recall from 65% to 90% ● Chatbot development using LLMs: Built the company’s primary chatbot from ground up utilizing OpenAI GPT APIs, prompt engineering. Also built another chatbot utilizing open source LLMs (MPT and Llama models) ● Speech recognition improvement: Improving the company’s speech recognition model to handle noise overlaid on speech. Built audio transcription pipelines utilizing speechSegmenter, ‘noise reduce’, and OpenAI whisper

DB Digital | Sep 2021 - May 2022

Senior Data Scientist

● Constructed an ensemble clustering model utilizing multilingual BERT and LaBSE to cluster Hindi news articles. This model enriched the content metadata, reduced manual workload for annotators by ~40%, and facilitated personalisation tasks ● Contributed to development of the model storage, model versioning components of the company's data science platform from scratch. These efforts helped create a strong foundation which eased scalable deployment of data science models

Disney + Hotstar | Apr 2019 - Jul 2021

Data Scientist

● Built a content recommendation system, using association rules algorithm, for notifications and customer reactivation initiatives. This resulted in improved customer retention, and a ~2% increase in customer reactivation ● Implemented end-to-end pipelines for key business metrics through ETLs (on spark and hadoop) for the huge 10TB+ per day viewer activity data. Used multiple tools including Airflow, Redash and AWS services like S3, EC2, EMR for effective setup ● Built an effective churn prediction model using logistic regression on spark to drive retention initiatives of viewers. Strategies resulting from this model led to ~3.5 % increment in retention ● Conceptualized and built customer segments to monitor and cater to the massive viewer base across India. These segments are continually utilized across the company for targeted marketing efforts

Flipkart | Sep 2017 - Apr 2019

Data Scientist

● Conducted extensive causal analysis and A/B tests to identify pain points in payments experience. By collaborating with design, product teams, proposed and implemented a personalized payments experience leading to ~2% increase in payment conversion ● Aided in conceptualization and implementation of product enhancements such as save for later, collated offer experience, Debit Card Equated Monthly Installments (EMI) which led to ~1.5% improvement in order conversion

Tredence | Feb 2016 - Sep 2017

Senior Business Analyst

● Applied a range of ML methods including regression analysis, time series forecasting, clustering to deliver analytics consulting projects in operations management, strategy, forecasting areas for US clients in CPG, FMCG, Industrial rentals, Retail sectors

My Education

Take a look at my formal education

Masters in Data ScienceUniversity Of San Francisco | 2023

Bachelors in Mechanical EngineeringBirla Institute of Technology and Science Pilani | 2015

About Me

ML and Gen AI professional with 7 years experience in OTT entertainment, News Media, management consulting, e-commerce industries. Worked on a variety of impactful business problems throughout my experience.

Create Your Data Portfolio for Free