Herbert George
22yo software engineer based in India, Kerala.
AI/ML Engineer | Deep Learning • Reinforcement Learning • Robotics
I build fun and cool things.
Care for a quick game?
Try my games :D
Wood N Beyond
Web Technologies Freelancer
- Designed and deployed a production-ready ReactJS website, improving client visibility by 40%.
- Optimized frontend workflows improving load times and retention.
Skills
Featured Projects
view more →
Reinforcement Learning for Genome Editing
A simulation-driven genome editing framework that models DNA sequences as dynamic environments. Reinforcement learning agents trained with DQN and PPO learn to perform codon-aware nucleotide edits, optimizing edit accuracy while minimizing off-target mutations. The system demonstrates how sequential decision-making can be applied to complex biological editing problems, achieving faster convergence and higher precision compared to heuristic methods.
MFIDS – Multi-Faceted Intruder Detection System
A smart, camera-based security platform , designed to be fully programmable and extensible. MFIDS uses real-time computer vision to detect and recognize people, triggering automated actions such as unlocking doors or activating connected devices for authorized users. When unknown individuals are detected, the system raises alerts and flags anomalies, making it adaptable to a wide range of security and access-control scenarios beyond just door locks.
Featured Research Papers
view more →Deep Learning-Based Drug Repurposing Using Knowledge Graph Embeddings and GraphRAG
Herbert George, et al.
bioRxiv · Dec 10, 2025
This paper presents a deep learning-based strategy for drug repurposing using knowledge graph embeddings and GraphRAG. By combining pretrained TransE model embeddings on the DRKG with LLMs, the system processes natural language queries to retrieve relevant drug candidates and generate interpretable explanations of drug-disease relationships.
Nova: A Hybrid Deep Learning and Conversational AI Framework for Fashion Recommendation
Herbert George, et al.
IEEE Preprint · TechRxiv · Dec 8, 2025
Nova is an AI-powered fashion recommendation system offering tailored outfit recommendations and an AI chatbot. Utilizing deep learning models like GPT, LSTM, and BERT alongside massive datasets, Nova combines NLP and machine vision to empower consumers with personalized, sustainable styling guidance.
Featured Certifications
view more →
Stanford Online
Issued Oct 2025
Advanced Learning Algorithms
ID: 1UR9ZQPF2O7W
Oracle
Issued Jul 2025
Oracle Cloud Infrastructure 2025 Certified Generative AI Professional
ID: OCI25GAIOCP