I'm Bob Kranson. I run Kranson Enterprises, where 30 years in tech goes into practical systems work: manufacturing information systems, Azure, DevOps, and AI pipelines. I share technology support, development, and community knowledge at MIJUG (My Information Just Under Glass).
Manufacturing information systems with Uniface. My current professional engagement is developing manufacturing information systems in Uniface. For professional background and experience, see my CV site.
Memory and boundaries for AI. "Whose Memory (for AI) Is It Anyway?" is my current focus: what AI assistants remember and who owns it.
- Better framing before build starts
- Repeatable patterns over heroics
- Consistency that scales across teams
Translation: fewer surprises, cleaner delivery, better outcomes.
Upstream and home lab. Three of my pull requests are merged into raspberrypi/rpi-image-gen, Raspberry Pi's own image builder (#234, #353, #365): I contributed its example network layers for ifupdown, Netplan, and systemd-networkd, and keep them building as the base evolves. The same builder powers ke-net-screen, a Raspberry Pi 5 home DNS appliance that pairs Pi-hole, Unbound, and Avahi to block ads and trackers for the whole house. The Little Box That Says No tells the story.
- ollama-embeddings: a Rust application that parses directory files and generates AI embeddings using the Ollama API.
- mijug-dot-net: the public repo behind mijug.net, open for content requests and security reports.
- gota-react: a full-stack sample pairing a Node.js project with a Go modules backend and a Jekyll documentation wiki.
- example-networks: Linux example network configurations, the standalone copy of my upstream contribution.
- rpi-image-gen (fork): the working fork those upstream pull requests come from.
- rpk-cv: my curriculum vitae and resumes, published at devopsbob.github.io/rpk-cv.
const devOpsBob = {
currentFocus: [
"Device and database integrations with Uniface",
// A wink to "Whose Line Is It Anyway?" - but this time, memory points matter.
"Whose Memory (for AI) Is It Anyway?",
],
shipped: [
"3 merged PRs in raspberrypi/rpi-image-gen",
"ke-net-screen, public and building",
"mijug.net v18 with cookie consent",
],
homeLab: "ke-net-screen: the Pi 5 that tells ads and trackers no",
nextUp: "A/B boot slots for ke-net-screen images",
learning: "Keeping AI memory local: per-workspace, out of git, and mine",
seeking: "Collaborators for open source Raspberry Pi and networking work",
needsHelp: "Building a steady pipeline of work for humans in the tech business",
};π¬ Ask me about:
- π¦ TypeScript's evolution and modern development
- π€ Uniface and AI pipeline integrations
- π’ Node.js adventures
- βοΈ Azure cloud deployments
- βοΈ DevOps best practices
- π Home lab setups
π¬ Reach out: LinkedIn or mijug.net | π CV: devopsbob.github.io/rpk-cv | π·οΈ Pronouns: He/Him
πΉοΈ Retro Tech Survivor: Been online since dial-up BBS days in the 1980s - yes, I remember 300 baud modems!
π Eternal Student: Decades in and still excited about every new framework, language, and paradigm shift
π« University Course Creator & Instructor: Created and taught a university course on designing intelligent information systems, connecting software design concepts with practical development.
π‘ Home Lab Enthusiast: Because why rent cloud when you can build your own datacenter in the garage?
The month-by-month journal for 2025 and 2026 lives in history.md.



