Hire Rihaz R. - RAG Engineer

* Achieved 90%+ faithfulness and context recall in RAG pipelines by implementing RAGAS-based validation and systematic retrieval tuning. * Improved document retrieval quality by identifying and eliminating near-duplicate content across 400+ PDFs using TF-IDF and SBERT embeddings, leading to more consistent grounding and reduced retrieval noise. * Designed and validated a CPU-only real-time vision pipeline under edge deployment constraints, achieving ~8.9 FPS while preserving stable multi-person tracking in live streams. * Diagnosed data leakage and class imbalance during exploratory analysis and introduced corrective feature engineering strategies, improving model robustness as measured by Precision, Recall, and F1-score.

* Analyzed a ~10,000 enterprise dataset to support identification, evaluation, and onboarding of 500+ enterprises, contributing to full target achievement across multiple development schemes * Built Excel-based dashboards (Pivot Tables, formulas) to track enterprise status, fund utilization, and scheme-wise performance for district-level monitoring.

* Designed and implemented a facial recognition-based attendance system using OpenCV, enabling automated image capture and CSV-based attendance logging for classroom environments. * Built a SMS spam detection model using TF-IDF and Scikit-learn, achieving ~96% precision in filtering malicious messages during offline evaluation. * Mentored students on academic and mini-projects, guiding data preprocessing, EDA, logic building, and result interpretation using Python.