Hi, I'm Vishnu Raj
Software Architect & DevSecOps Engineer

Know more

About me

Profile Image

I'm a Software Architect and DevSecOps Engineer with expertise in designing and deploying secure, scalable infrastructures and AI-driven applications. Currently conducting research at Technical University of Ilmenau on cross-modal image registration and semantic segmentation.

My specializations include backend development with FastAPI, AI/ML integration (LLMs, RAG, WhisperX, Pyannote), and cloud-native deployments using Docker, Terraform, and major cloud providers. I'm passionate about building secure, observable systems aligned with GDPR, ISO/IEC 27001, and EU AI Act standards.

View Resume

Projects

Cross-Modal Image Registration & Semantic Segmentation

Engineering cross-modal image registration pipelines for high-dimensional sensor data (RGB, MWIR, 5D) using Fourier-based phase correlation and SIFT/ORB descriptors. Optimizing CNNs for real-time semantic segmentation and automating data preprocessing workflows for multi-terabyte industrial datasets.

AI-Powered Contract Analysis Platform

Architected multi-cloud infrastructure on Hetzner and GCP using Terraform. Developed FastAPI backend with Celery task queues, JWT auth, and GDPR-compliant data flows. Implemented RAG pipelines with Qdrant vector search and open-source LLMs for real-time contract interpretation with Prometheus, Grafana, and Loki observability stacks.

Enterprise SSO Integration & Portal Development

Designed and deployed enterprise portal with ADFS/Exchange SSO integration. Implemented secure authentication with Active Directory and on-premises services. Managed infrastructure spanning Debian servers, Windows AD, and Exchange Server integration for seamless unified access management.

Hybrid ML Model for Energy Theft Detection

Implemented hybrid deep learning architecture combining CNNs and Multi-Layer Perceptrons (MLPs) for energy theft detection. Engineered preprocessing pipelines converting time-series data into daily and weekly representations for improved model generalization and accuracy.

Get In Touch

Feel free to reach out!

Get in Touch