ashish ranjan

Designed and implemented a continuous learning pipeline using Azure services for automated object detection model evaluation, comparing with previous models, and triggering retraining when necessary. Fine-tuned object detection models based on real-world data, significantly improving accuracy and ensuring only high-performing models were deployed.

Developed an AI-driven, RAG-based test case generator that retrieves, refines, and structures test cases from transaction data, regulatory frameworks, and banking SRS documents. Implemented constraint-based test case filtering to align with financial regulations.

Developed a video surveillance system for human and crowd activity detection, optimizing models for accuracy-speed trade-off and analyzing complex crowd behavior patterns. Fine-tuned multiple classification, detection, segmentation, and transformer models, and designed a user-centric interface.

Completed a Bachelor of Technology in Computer Engineering, building a foundational understanding of computer science principles. Gained expertise in various aspects of computer engineering.

Pursued a Master of Technology in Computer Science, focusing on advanced topics in the field. Engaged in research and development activities related to deep learning.