SYSTEMS · ARCHITECTURE · WORKFORCE TECHNOLOGY
Nobody calls these software problems. I've spent my career learning that they usually are. Somewhere between airlines, logistics, healthcare, and the workforce systems behind hospitality, I landed on two convictions I can't shake: architecture is employee experience, and great systems only come from teams that trust each other. So I build both. Platforms that absorb complexity before it reaches a human, and engineers who grow into leaders.
Most recognition platforms measure scheduling luck and call it merit. Fairness has to be modeled: actual performance separated from opportunity to perform. I wrote about the data model this requires.
Legacy isn't just technical debt. It's human debt: the side spreadsheets and tribal knowledge people build around old systems. Modernize without understanding those behaviors and you ship the same friction in a newer interface.
When enterprise software is fragmented, the pressure moves to managers and frontline staff. I judge architecture by how much complexity it absorbs before a user ever sees a screen.
Amber marks friction flowing toward a frontline human. Flip the architecture on, or press and drag (or swipe) to absorb particles yourself.
Complexity absorbed: 0
Reached a human: 0
Peer-reviewed work on what changes when the component you're shipping is probabilistic, and the editorial boards where I review other people's answers to the same question.
Prompt injection is a systems-engineering problem, not a model problem. Language models flatten instructions and data into one stream, so you can't instruct your way out. I lay out a containment architecture: least-privilege access, dual-model designs, constrained outputs, human oversight, and continuous monitoring.
Hampton Global Business Review ↗Fifty years of software assumptions (deterministic, obedient, nearly free to scale) don't hold for probabilistic systems. I propose five operational boundaries for trustworthy AI: what it can reach, what you will believe, what you can reproduce, what it costs to serve, and how much of it you actually build.
Global Business & Economics Journal ↗A profile on the practical evolution of workforce technology: why data flow, integration reliability, and absorbed complexity decide the employee experience.
Success Quarterly ↗On building enterprise systems: human debt, ratio-based recognition fairness, and why DORA metrics belong in the same conversation as human impact.
HackerNoon ↗The argument underneath all of it: Architecture Is Employee Experience — a manifesto.
I'm Director of Engineering at Unifocus, where I own the Workforce Performance product line: planning and scheduling, time and attendance, budgeting, and the integrations that connect them to the systems hotels already run. Before this I built for airline travel platforms, logistics, and healthcare, environments where downtime is a grounded flight or a delayed patient, not a ticket. I lead global teams across the US and India, I'm accountable for uptime, client success, and budget, and I measure myself as much by the leaders my teams produce as by what we ship. I also serve as Associate Editor at the Global Business & Economics Journal and the Hampton Global Business Review, where I cover enterprise SaaS architecture, technology governance, and AI-driven operations.
0+
Years in enterprise software
0
Industries built for
0
Global team hubs
0
Editorial boards
Anil Kumar Karumuru is Director of Engineering at Unifocus, where he leads workforce performance platforms for the hospitality industry. His nearly two decades in enterprise software span airline technology, logistics, healthcare innovation, and workforce management. He serves as Associate Editor at the Global Business & Economics Journal and the Hampton Global Business Review, and his research on trustworthy AI engineering and prompt-injection resilience appears in both. He writes about modernization, integration architecture, and fairness in workforce systems, with work appearing in Success Quarterly, HackerNoon, and Medium.
Building something in workforce tech, or want to compare notes on modernization?