Vedika Darokar

Worked on developing real-time AI systems for clinical workflows, including speech recognition and structured data extraction. Built end-to-end ASR pipelines using Faster-Whisper and integrated LLM-based prompt engineering to generate structured, schema-compliant outputs. Designed a voice-driven time-out protocol system for clinical workflow validation and data capture. Developed API-based inference services using FastAPI and Docker, and optimized GPU-based pipelines for low-latency performance. Skills: Sp

-Led the end-to-end project to implement an automated dock management system, integrating real-time data with AGV technology. Coordinated with cross-functional teams to streamline operations, achieving a 20% increase in efficiency. -Optimised machinery operations, reducing maintenance costs by 10% and enhancing reliability. -Developed and applied 12+ machine learning algorithms, including Random Forest, SVM, K-Means, PCA, and Neural Networks, using advanced visualisation tools like Matplotlib to extract act

- Leveraged Python libraries to conduct in-depth data analysis on heavy- duty vehicles, analysing over 10,000 data points. - Identified key performance trends with 95% accuracy, enabling improved decision-making and efficiency. - Implemented a variety of Machine Learning algorithms and employed sophisticated visualisation techniques to interpret data.

Specialisation- Artificial Intelligence and Autonomous Vehicles and Connectivity