Hi, I'm Ali Ahmadi Esfidi

A project launched by the NORC Lab, focused on developing an application for diagnosing bridge damage. Worked as a ReactJS and Express.js developer and contributed to the design and implementation of computer vision models for damage detection.

Collaborated on implementing DRL algorithms (DDPG, PPO, A2C, D3QN) to optimize irrigation policies. Designed a custom Gym-like environment, integrated DSSAT-generated datasets, and conducted comparative analysis between different RL models.

Builds upon the KH-99 model and introduces a novel Gap-Bracket CYK parsing algorithm to improve prediction accuracy, especially for complex motifs like pseudoknots. Preliminary version presented at the CBRC Journal Club.

Developed, evaluated, and interpreted machine learning models for classifying gold-binding peptides. Applied various ML algorithms and feature engineering techniques to achieve high classification accuracy.

Cross-platform desktop application built with Electron and TailwindCSS for efficient image management, annotation, cropping, and segmentation, targeting machine learning datasets.

Implemented and analyzed Orthogonal Gradient Descent (OGD) to mitigate catastrophic forgetting in deep neural networks during continual learning scenarios.

Full-stack productivity tool for organizing daily tasks, reminders, and expenses, built with React (component-based) frontend and Express.js/MongoDB backend.