Moein Bagheri

• GPA 17.26/20 overall · 18.5/20 over the last three semesters · 18.5/20 in specialized courses
• Major track: Artificial Intelligence · Minor track: Information Technology
• Selected courses: Signals & Systems (20) · Language & Speech Processing (20) · Design of Algorithms (20) · Data Mining (19.75) · Probability & Statistics (19.25) · Applied Linear Algebra (18.87) · Deep Learning (18.5) · Information Retrieval (18.25) · Artificial Intelligence (18)
• Final-year project (in progress): Persian–English code-switched speech recognition · Supervisor: Dr. S. F. Noorbehbahani

Admitted through the national entrance exam for Iran's schools for exceptional talents

• Built plain and residual 1D CNNs on MFCC features to classify four emotions in an IEMOCAP subset.
• Compared random and speaker-independent splits across six experiments: accuracy fell from 65.1% to 58.0% on unseen speakers, showing how random splits overstate generalization.
• Released the trained checkpoints and a command-line predictor for WAV files.

• Built a fully local question-answering system over four retrieval papers (REALM, DPR, RAG, FiD) that cites the PDF pages it draws on.
• Routed requests to six research tools, with a keyword fallback when the model returns an invalid tool call.

• Built a Persian search engine with Hazm preprocessing, an inverted index, Boolean retrieval and TF-IDF/BM25 ranking.

• Modeled melodies as note-token sequences with unigram to trigram models, Lidstone smoothing tuned by leave-one-out cross-validation, and perplexity evaluation.

• Trained GRU networks for five-class heartbeat classification (MIT-BIH) and normal/abnormal classification (PTBDB).

• Extended the competitor analysis and ran user interviews for a type-2 diabetes self-care concept.