Our Blog
Explore Insights, Trends, and Expertise
Welcome to the Rational IT Solutions Blog! Here, we share insights, tips, and case studies on topics ranging from AI and ML to digital transformation and sustainable architecture. Explore our latest articles to stay updated on industry trends and best practices.
Agentic AI with RAG: Building Workflows That Think, Act, and Improve
Agentic AI is a new way of building software: instead of returning a single answer, LLMs plan multi-step processes, execute them iteratively, and improve through reflection and tool use. Grounded by RAG, that becomes a workflow you can trust.
Read article →
When Design Thinking Meets Hypervelocity Engineering
Design thinking tells you what is worth building; Hypervelocity Engineering (HVE) and its open-source HVE Core framework tell you how to build it in weeks instead of years; Forward Deployed Engineers make it real inside the client's world. Here's how the three compose.
Read article →
Cloud Storage for the AI Era: Multi-Cloud and Hybrid Object Storage with MinIO
AI workloads live and die by their data layer. Here's how MinIO's high-performance, Kubernetes-native object storage anchors a multi-cloud and hybrid strategy — one S3-compatible API across every environment you run.
Read article →
The Six Building Blocks of Multi-Agent AI Systems
Agentic AI demos are easy; production multi-agent systems are engineering. Six disciplines decide whether a fleet of agents is dependable or dangerous: routing, security, authentication, authorization, offline operation, and service orchestration.
Read article →
From DevOps to AgenticOps: MLOps, LLMOps, AIOps — and Least-Privilege Security at Every Layer
Plan, code, build, test, release, deploy, operate, monitor — the DevOps loop now has descendants. A tour of MLOps, LLMOps, AIOps, and AgenticOps, and the least-privilege security principle every one of them depends on.
Read article →
@RitS Labs: Innovating with AI and Quantum Technology Stacks
@RitS Labs is our engine for innovation: hands-on experimentation with AI and quantum technology stacks, run in collaboration with intern and summer programs, and matured into client-ready solutions through a disciplined lab lifecycle.
Read article →
Unlocking Data Potential with Gen AI Multi-Agent Tooling: The RAG Framework
How Gen AI multi-agent tooling with a Retrieval-Augmented Generation (RAG) framework streamlines data ingestion, retrieval, and insights generation across structured and unstructured datasets.
Read article →More from Medium
Deeper technical write-ups by our founder
Building a MiniGPT from Scratch with JAX: Architecture, Training, and Inference on Mac M1 Pro
A 20-million-parameter GPT-2-style transformer built from first principles on a MacBook — architecture, data pipeline via Grain, JIT-compiled training, Orbax checkpointing, and a Gradio chat interface, using the same JAX stack that powers Gemini.
Read on Medium ↗Running Claude Code Locally with Ollama — CLI & VS Code Setup: What Actually Works (and What Doesn't)
A practical guide to running Claude Code fully offline with Ollama and IBM's Granite 3.3 model — covering both the terminal CLI and the VS Code extension, and why native tool calling is non-negotiable for it to work.
Read on Medium ↗