Independent Researcher (ORCID: 0009-0009-5758-3845)

Edge AI Systems Architect | Distributed Intelligence | Applied AI Research

I design and prototype computationally efficient artificial intelligence systems for resource-constrained edge environments. My work focuses on bridging high-complexity machine learning models with low-latency, energy-aware execution architectures used in healthcare, environmental monitoring, and distributed sensing systems.

My portfolio combines systems engineering, applied machine learning, and research exploration, focusing on making advanced AI models practical for decentralized real-world infrastructure.

Repository https://github.com/swapins/gnn-edge-systems-analysis

A systems-level investigation into deploying Graph Neural Networks on edge hardware for biological network analysis.

The project explores:

• Graph Neural Network inference under hardware constraints
• latency and memory behavior of GNN architectures
• biological interaction network modeling
• architectural strategies for hardware-aware machine learning systems

This repository serves as an experimental framework for studying Edge-native GNN deployment.

My research interests sit at the intersection of artificial intelligence systems, graph neural networks, and computational biology.

Research areas include:

• Graph Neural Networks (GNN)
• Edge AI systems
• Computational biology
• Protein interaction network modeling
• Decentralized machine learning systems
• Hardware-aware neural architecture design

Academic researcher identifier:

ORCID https://orcid.org/0009-0009-5758-3845

A decentralized anomaly detection framework designed for real-time physiological signal analysis on single-board computers.

Focus areas include:

• signal normalization stability
• low-latency inference
• fault-tolerant distributed monitoring architectures

Experimental framework for Graph Neural Network architectures applied to protein-protein interaction networks.

Research focus:

• normalized graph convolution stability
• sparse graph processing
• feasibility of GNN inference on constrained hardware platforms

A modular experimentation platform integrating:

• computer vision detection models
• IoT telemetry streams
• GPS-linked environmental datasets

Designed to evaluate distributed AI inference strategies in low-connectivity agricultural environments.

Experimental edge architecture combining environmental sensor fusion with AI-assisted biological indicator detection for ecological monitoring research.

• PyTorch
• TensorFlow Lite
• Graph Neural Networks
• Model quantization and optimization
• NVIDIA Jetson and SBC-class hardware

• PHP (Senior)
• Laravel
• PostgreSQL / MySQL
• Redis
• RESTful and GraphQL APIs

• Docker
• Kubernetes
• Linux systems administration
• MQTT messaging systems
• Event-driven microservices

• Protein-protein interaction networks
• TCGA dataset preprocessing
• Healthcare interoperability concepts (FHIR)
• Privacy-aware distributed systems

• MBBS Coursework — Government Medical College, Thiruvananthapuram
• Diploma in Business Administration (In Progress)

My work integrates medical domain understanding with systems engineering to explore scalable decentralized AI architectures.

• Hardware-aware Graph Neural Network optimization
• Stability-aware neural model compression
• Energy-efficient edge inference scheduling
• Distributed anomaly detection across edge nodes
• Edge-native AI architecture design

Singapore-based.

Open to research collaboration, interdisciplinary AI systems development, and systems-level architecture discussions.