Ayush Ojha

Designed and implemented an end-to-end OCR + NLP research pipeline in Python, processing 10,000+ handwritten manuscript pages with 85% accuracy using custom preprocessing algorithms and deep learning models Engineered an efficient entity recognition and relation extraction system, applying algorithmic optimizations to unstructured text parsing — improving extraction accuracy by 40% Built a searchable database system with optimized indexing and retrieval algorithms, reducing query response time by 60% through efficient data structure design

AI and ML Intern - RMSO Services Private Limited - Lucknow, India

(2025-05 - 2026-05)

• Designed and implemented an end-to-end OCR + NLP research pipeline in Python, processing 10,000+ handwritten manuscript pages with 85% accuracy using custom preprocessing algorithms and deep learning models
• Engineered an efficient entity recognition and relation extraction system, applying algorithmic optimizations to unstructured text parsing — improving extraction accuracy by 40%
• Built a searchable database system with optimized indexing and retrieval algorithms, reducing query response time by 60% through efficient data structure design
• Optimized model performance through systematic hyperparameter tuning and experimentation, achieving 30% faster inference while maintaining accuracy for production deployment

AI Intern - Microsoft, SAP, AICTE & Edunet (Virtual) - Remote

(2024-12 - 2025-01)

• Collaborated with 5 industry experts on research-driven model development, integrating new architectures into automated preprocessing and deployment pipelines — improving efficiency by 15%
• Enhanced Human Pose Estimation model accuracy from 75% to 90% by implementing advanced CNN architectures and systematic data augmentation experiments, enabling real-time inference
• Conducted rigorous model evaluation using precision, recall, and F1-score; documented findings and implemented iterative improvements based on empirical performance analysis

Digital Marketing Intern - A2Z Journals, Global Education & Research Institute (Virtual) - Remote

(2023-01 - 2023-07)

• Designed data-driven, segmentation-based campaign experiments — open rates ↑25%, CTR ↑15%, unsubscribe rates ↓12% through structured behavioral analysis

B.Tech - Artificial Intelligence - Amity University Lucknow (2022-09 - 2026-06)