Apr 2023 – May 2024 · 1 yr 1 mo

• Collaborating with the Sprinklr Insights Backend Team, I played a pivotal role in enhancing the platform through the implementation of impactful features and resilience enhancements for the module.
• Architected a highly extensible and scalable unified adoption tracking framework for Sprinklr, leveraging Java, Elasticsearch, Kafka, and Webhooks. This framework provides a consolidated view for monitoring diverse client feature adoption metrics, streamlining the tracking process.
• Revamped the backend infrastructure of several resource-intensive pages within Sprinklr's user interface, significantly optimizing load times from over 30 seconds to an impressive 5-6 seconds.
• Engineered an asynchronous enrichment framework, utilizing multithreading concepts and Kafka, to enhance Sprinklr's new survey management module. This framework, focusing on sentiment analysis and word cloud generation, adds substantial value to customer survey text data analysis, contributing to a more comprehensive understanding for users.

Jul 2021 – Mar 2023 · 1 yr 8 mos

Jan 2021 – Jun 2021 · 5 mos · Bangalore Urban, Karnataka, India

May 2020 – Jun 2020 · 1 mo · Gurgaon, Haryana, India

Jan 2019 – Apr 2019 · 3 mos · Pilani, Rajasthan

• Worked under the guidance of Dr. Dhiraj Sangwan at CSIR - Central Electronics Engineering Research Institute, Pilani(C.E.E.R.I). GAN refined images were used for training a two-stage network comprising of the region proposal network followed by an image classification network was used. A different non-max suppression function from the FASTER-RCNN network was used to merge boxes using the overlap and not IOU as the objects to be detected were very small. The accuracy obtained was far better than only FASTER-RCNN based network.

May 2018 – Jul 2018 · 2 mos · Thiruvananthapuram Area, India

• Project was titled "AI Based bug Tracking", basically I had to group various tester/user reviews into a specific number of predefined defects/bugs which would further help in root cause analysis of the defect. Used various Natural Language Processing Techniques like Word2Vec, FastText, GloVe, for generating word embeddings.
• Sentimental analysis on user reviews for analysis on the severity of app crash/defect.
• Multi Feature Text Classification using pipelines.
• Web Scraping to generate domain specific corpuses for word embedding training.
• Used Flask for deploying the code.