S Karthikeyan

Developed and implemented a clustering model to categorize users based on key features related to invoice processing, resulting in a 20% improvement in business efficiency. Identified patterns causing delays in invoices, enabling targeted interventions and contributing to significant process enhancements.

Designed and tested an auto-labeling model using a pre-trained cross-encoder NLI model for clustered tickets which reduced the time taken to label the tickets by 25% and improved the quality of the data passed to the model by 30%. Engaged in Federated Learning to create a super model incorporating insights from previous customer data. Utilizing transfer learning to apply this knowledge seamlessly to new clients. Contributor in Developing a ticket Queuing system which assign ticket the best service engineer based on SLA time and wait time worked majorly on System designing of the feedback system from the ops queue.

Pioneered the development of an AutoML solution, enabling users to effortlessly explore and comprehend data without requiring a background in data science. Established a streamlined pipeline for fundamental machine learning features, allowing a systematic approach to model building. Adapted models based on data size, enhancing efficiency in the model definition process.