I am the developer of FUNIN, a web platform designed to organize and provide access to GH32 fungal enzyme data. The system was built with Python, Flask, MongoDB, and Docker, and serves as a foundational dataset for future AI-driven enzyme design.
Currently, my research focuses on the optimization of GH32 fungal enzymes derived from the FUNIN database, using Protein Language Models (pLMs) and deep learning frameworks such as PyTorch and TensorFlow.
My research interests include Protein Language Models (pLMs), Large Language Models (LLMs), Machine Learning, Genetic Algorithms, and Protein Design & Engineering.
Tech stack: Python • Flask • MongoDB • Docker FUNIN is a web application designed to catalog and visualize GH32 fungal inulinases relevant to bioethanol production. It integrates curated enzymatic and sequence data to support AI-based enzyme engineering.
Publication: DOI: 10.1016/j.ijbiomac.2025.145050
Tech stack: Python • Docker • MongoDB • CLEAN (Enzyme function prediction using contrastive learning) A comprehensive ELT (Extract, Load, Transform) workflow that extracts data from InterPro and UniProt databases to build the FUNIN database. The pipeline integrates multiple data sources and implements enzyme function prediction using contrastive learning approaches for comprehensive enzyme characterization.
Tech stack: Python • Scikit-learn A predictive software for evaluating antimicrobial peptide activity against SARS-CoV-2.
- Current Focus: Optimization of fungal GH32 enzymes using ESM-based pLMs, genetic algorithms, and deep learning architectures for protein function prediction and engineering.
