Akash Kashyap

Co-authored and published a peer-reviewed AI system for automated resume screening and job recommendation in an indexed international journal. Designed an NLP pipeline using TF-IDF vectorization, Cosine Similarity, and Named Entity Recognition (NER) to extract skills, experience, and qualifications from PDF resumes. Trained and benchmarked SVM vs. Logistic Regression classifiers; SVM achieved superior accuracy in high-dimensional resume-to-job description matching. Built a ranked job recommendation engine using similarity scoring for real-time, data-driven candidate-to-role matching.

Engineered a real-time road defect detection pipeline using YOLOv8 and OpenCV, achieving 92% classification accuracy on a held-out test dataset. Deployed a Flask web application for end-to-end video upload, inference, and annotated output rendering; reduced manual road inspection effort by 80%.

Architected a multi-source threat intelligence platform integrating VirusTotal, AbuseIPDB, and Shodan APIs for automated IP reputation scoring and file-hash risk analysis. Built a Flask REST backend with psutil-based live system monitoring and SQLite for persistent audit-ready threat logs. Delivered Power BI dashboards for threat visualization; reduced manual analyst monitoring effort by 60%.