natrual-language-processing

The project focuses on sequence labeling for Vietnamese text using three classical and neural approaches: Hidden Markov Model (HMM), Conditional Random Fields (CRF), and BiLSTM-CRF.

Homework assignments for CMU 11-611 Natural Language Processing (Spring 2026) — covering language identification, n-gram LMs, text classification, machine translation evaluation, and DPO fine-tuning.

Maglev is a Rails-native, read-only knowledge and query layer for ActiveRecord applications.

Production-ready Financial NLP Pipeline: Fine-tuning FinBERT with LoRA (PEFT) for 98% accuracy. Features automated evaluation & Power BI observability dashboard.

Language Modeling & Spelling Correction

Comparative NER study: Regex vs spaCy vs DSPy LLM. Measures precision, recall, F1, cost, and latency across synthetically generated records with Streamlit dashboard.

AI-powered intent extraction library that converts natrual language input into structured developer-defined data.

Enhancing Document-Level Relation Extraction with Anaphor Nodes and Visual Transformation

AI in Banking Application showcases how artificial intelligence improves banking services through fraud detection, customer support, risk analysis, and automated decision-making

Generates plain-language narratives from R statistical objects, model output, ggplot figures, and datasets, via any LLM that the `ellmer` package supports.

this project implements a python-based solution to detect and redact PII and SPI from text documents, PDF's, and images.