Swati Gupta

Established robust collaborations with external stakeholders to design and streamline data generation processes. Innovated methodologies for creating comprehensive datasets tailored to translation requirements, facilitating the organization's mission to combat human trafficking on a global scale. Coordinated the gathering and preprocessing of extensive multilingual datasets, encompassing Spanish, through the application of web scraping techniques. Transformed raw data into structured formats optimized for model ingestion. Led the comparative analysis of baseline models. Executed a Proof of Concept (POC) on fine-tuning ML models for Spanish to English translation, focusing on MHSHT-related news articles. Continuously refined models to ensure real-time monitoring and reporting capabilities. Utilized SacreBleu for rigorous evaluation. Provided mentorship and constructive feedback to junior team members, fostering a collaborative environment conducive to skill development and knowledge sharing.

Implemented a Classification ML model to categorize sectors within the data obtained through web scraping. This enhanced data analysis and provided valuable insights for decision-making processes. Achieved an impressive total of approximately 1,460 hours of time savings in a year, leading to increased efficiency, cost savings, and faster decision-making processes. Conducted analysis of technological extracting and processing webinar related information from multiple sources using Web Scraping, REST API within Flask Microservice Framework. Applied NLP on top of it using a Knowledge dictionary, to get the insights to share with marketing teams.

IntelliAssure is a Wipro IP having ML Classifier to Categorize Defect Logs using Machine Learning Classification techniques like Logistic Regression, Random Forest, SVM. Reduced manual human efforts by 18%, resulting in significant time and cost savings. Conducted rigorous experimentation to optimize the performance of the algorithms.

Developed a Python script which resulted in enhanced workforce management, enabling the matching of future demand and planned ramp downs based on specified band criteria. Empowered decision-makers with accurate skill requirement forecasting and demand analysis 30 to 90 days in advance, leveraging historical data and understanding influencing factors. Facilitated strategic resource allocation and optimized workforce planning, enabling proactive matching of resources for future demands and planned ramp downs.

Developed a secure, scalable payment web app using Python, Django, and AWS, enabling money transfers, request handling, and multi-currency transactions with robust security measures. Also developed a project empowering Sign Language Communication Through Image Based Recognition, Domain Specific Language Modelling.