Hire Devadarshini J. - Fine-Tuning Engineer

I led the design and development of ML-driven recruitment intelligence systems, including recruiter performance and analytics dashboards used by 30+ recruiters for real-time tracking of mandates, productivity, and closures. I improved the resume parsing pipeline by increasing model accuracy from 75% to 87% through preprocessing optimization, model fine-tuning, and validation improvements, and built a card-parsing tool for structured candidate data extraction. By automating reporting workflows and integrating analytics pipelines into the product, I reduced manual reporting effort by 86% while contributing to product-level architecture decisions and ML system optimization.

* Developed AI-driven automation pipelines for storyboard generation, object detection, and content extraction using Computer Vision and NLP. * Built and deployed a Flask-based web application that automates extraction of Learning Objectives (LO), Standard Operating Procedures (SOP), and Storyboards (SB) from videos, images, and documents, with seamless PowerPoint integration. * Engineered a robust video segmentation pipeline for scene detection and AR-based storyboard creation, reducing manual effort by 70%. * Integrated Whisper ASR and Coqui TTS to enable multilingual narration, achieving 85%+ transcription accuracy even in noisy industrial environments. * Optimized deep learning models for object detection, image labeling, and scene understanding to improve efficiency in industrial training workflows. * Spearheaded end-to-end automation for training content creation by combining CV, NLP, and backend engineering for scalable deployment.