Let's Talk Business
Scroll down
About me
2021- Present
Started as the advisory and delivery partner for companies stuck between an AI pilot and a working system.
2026- Present
Built the intelligence layer those same companies needed: AI that connects to core legacy systems without replacing them.
A book on AI Orchestration: The Definitive Guide & a Research paper on Darshana: Six-School Framework for LLM Orchestration
Since 2021 I've been building Hundred Solutions. In 2026, Dversi: an intelligence layer that connects to the legacy business systems companies already run, instead of replacing them. It's built on MCP, and it compounds on what already exists rather than starting over.
Two decades in enterprise technology, including more than a decade leading large-scale IT services delivery across Scandinavia and Europe. Before founding Hundred Solutions, I spent years inside the ERPs, CRMs, and accounting platforms that AI is now supposed to orchestrate. I know why these systems break because I maintained them. That depth is the product, not the background.
"Intelligence built for the systems your business already runs on."
Work
Intelligence Layer
The intelligence layer for business systems. Dversi connects to your stack over the Model Context Protocol, so nothing gets replaced. It adds AI coworkers, workflows, and compliance enforcement on top of what you already run. Built to assist and support legacy systems, no rip and replace.
Customer’s existing systems preserved
Proven in accounting and insurance
Entity-centric knowledge graph
Advisory & Consulting
The activation and services partner for organisations adopting AI. We handle the scoping, integration, custom UI, and compliance work that turns the platform into a working system for a specific team, in eight to nine weeks, not eighteen months.
Scoping and integration into your existing stack
Custom UI built around one team’s workflow
Compliance work handled end to end
Publishes
Most organisations bolt AI onto legacy tools without understanding why the process exists. This book is a diagnostic before it's an implementation. Every framework in it was earned through production, every warning through failure.
Solve First, Automate Later. Zero-Token Workflows for the parts of a process that never needed a model in the first place. A multi-agent architecture that holds up in production instead of degrading. A framework for what makes an AI business defensible, not just demoable. None of it is theory, it's what building Dversi taught me.
Read on Amazon
Paper
Six schools of Indian epistemology, mapped onto the layers of an LLM pipeline: classification, retrieval, routing, ontology, synthesis. The finding that surprised even me: the same technique swings from 0% to 82% success depending on where in the stack it's applied. Tested across two open-source repositories, roughly five thousand generations, cross-judged by GPT-4o against generic controls.
Read the paper85%
Vritti: Epistemic self-classification
60% cross-judge (GPT-4o)
82%
Mimamsa: Query rewriting (6 Lingas)
67% cross-judge
93%
Nyaya: Tool routing via 4 Pramanas
70% search reduction
71%
Vaisheshika: Knowledge ontology (7 Padarthas)
vs generic entity ontology
82%
Vedanta: Output synthesis
63% cross-judge
Reads
Speaking
Most of my speaking happens in rooms where a decision is being made: a board evaluating an AI investment, a leadership team committing to a direction, a technical group designing something they will live with for years.
Working sessions with boards and leadership teams deciding where AI actually fits, and where it doesn’t.
The Definitive Guide to building automation that actually works. Read it before we ever get on a call.
Available on AmazonSolve the real problem first, manually if needed. Only once it's proven do we encode it into automation.
A framework session applying six schools of Indian philosophy to how your LLM systems are structured.