AI Strategy & Adoption
Nahar EmetSystems Architecture
I design and implement AI-powered business systems
and automation solutions that actually work.
Most AI projects fail because they chase technology instead of solving real operational problems. I help founders bridge strategy, operations, and technology — designing production AI systems for information retrieval, knowledge management, document processing, and workflow automation.
Whether it's building multi-service AI applications on containerised infrastructure, designing human-in-the-loop review systems for financial workflows, or helping an organization become AI-native — my focus is measurable business outcomes, not technology for its own sake.
Expertise
The technologies above are tools. They matter less than understanding how an organization actually operates — where information flows, where decisions get stuck, where automation creates relief instead of overhead. Most AI projects fail because they prioritize technology over operations.
Technical
Business & Leadership
Real experience building and scaling a business, designing production AI systems, and bridging strategy with technical execution.
Projects & Experiments
Client work is confidential. Here are the things I build in the open.
dot-prompt
Open Source
Prompt composition language and tooling. Write structured .prompt files with branching, variables, and fragments. See exactly what your AI receives before you send it. Open source under Apache 2.0.
ACS (Agent Coordination System)
Opening soon
A lightweight project manager for AI agents. Built on the Anantha app server. Agents claim tasks, lock files before editing, save and search memories, and sleep until dispatched. Exposed as ~30 MCP tools that agents call via function calls.
Prevents two agents editing the same file, duplicate work, and forgetting institutional knowledge. Does not manage conversation context, orchestrate multi-agent plans, or enforce anything — it is opt-in coordination, not mandatory supervision.
Opening soon on GitHub
Writing
AI in production is not about what models can do. It is about how information flows, how systems fail, and how architecture decisions compound. These essays are about that distinction.
Order, Eat, Toss↗Founder-Business Model Fit↗Budget AI Coding: 3 Billion Tokens for $20↗Why Workflows Beat Autonomous Agents↗Why Context Quality Determines AI System Quality↗Overengineering Destroys Startup Execution↗AI Systems Are Coordination Systems↗Operational Intelligence vs. AI Hype: What Actually Works↗
Let's build something.
Whether you need AI infrastructure, workflow automation, or just a conversation about how systems thinking applies to your business.