I am an MLOps-focused software engineer building end-to-end AI systems with deep learning and cloud-native architectures.
I have experience designing reproducible machine learning pipelines, experiment tracking workflows, model versioning, and CI/CD processes for deployable AI services.
I developed RetinaXAI, an end-to-end diabetic retinopathy screening platform using EfficientNet-B3, clinical-grade evaluation metrics, Grad-CAM explanations, and a RAG-based LLM reporting system.
I work across the ML lifecycle, from data ingestion, preprocessing, and model training to containerized deployment with Docker, Kubernetes, MLflow, DVC, and GitHub Actions.
I also have full-stack engineering experience with Next.js, React, NestJS, PostgreSQL, Prisma, Redis, and role-based access control systems.
I enjoy solving complex technical problems, leading collaborative teams, and applying AI and software engineering to impactful real-world products.
Built a multi-tenant e-learning platform using Next.js 15, React 19, NestJS, PostgreSQL, Prisma, Redis, and Turborepo.
Implemented hierarchical role-based access control for admin, agent, student, and parent portals. Delivered production deployments with health checks through GitHub Actions CI/CD.
Built RetinaXAI, an end-to-end diabetic retinopathy screening system using more than 35,000 EyePACS retinal fundus images, from data ingestion through clinical deployment. Trained an EfficientNet-B3 model for five-class severity grading, achieving 0.724 QWK, 78.7% accuracy, and 0.867 macro AUC on 7,022 test images.
Designed a RAG-based LLM system for structured clinical report generation and implemented a full MLOps pipeline with MLflow, DVC, Docker, GitHub Actions, and Kubernetes. Used mixed precision, gradient checkpointing, and gradient accumulation to train under GPU memory constraints, while validating Grad-CAM explanations with ophthalmologists.
Developed a virtual reality simulation in Unreal Engine 4 for psychological violence awareness training.
Implemented real-time behavioral tracking with Firebase, built a Scikit-learn vulnerability-scoring model, and created a Flutter dashboard for psychological analysis and decision support.
Led a 40-member team delivering social entrepreneurship initiatives.
Mentored students in agile execution and product development, and launched CaffeFree, a sustainable low-caffeine initiative.
Software Engineering, Software Engineering
IT Development Diploma, Information Technology Development
This professional hasn’t added portfolio projects yet.
This professional hasn’t listed any services yet.
