From the field

Building agentic systems, in practice

Field notes on the protocols, architectures, and engineering patterns we use to ship enterprise AI products on Alis Build. Concrete code, real diagrams, no hype.

Protocols9 min read

Introduction to MCP - the Model Context Protocol

How MCP turns "LLM with tools" from bespoke glue code into a stable, language-agnostic contract you can build a platform on.

Mar 25, 2026 Read article

Protocols11 min read

Introduction to A2A - agent-to-agent over gRPC

The Agent-to-Agent protocol gives multi-agent systems the same transport story REST gave web services. A walkthrough with Go.

Apr 9, 2026 Read article

AI Engineering10 min read

Agent search - wiring retrieval into an LLM agent

Why "search" isn't one tool but three layers, and how to expose them to an agent via gRPC tools without leaking implementation details.

Apr 23, 2026 Read article

AI Engineering10 min read

Generative UI - letting Gemini 3 design the screen

When the model returns a component tree instead of a paragraph, the agent stops being a chatbot and starts being a designer. The plumbing, in Go and Vue.

May 13, 2026 Read article

AI Engineering10 min read

Swapping the brain - building agents that aren't married to one model

The reasoning model is the most replaceable part of an agent. How we keep it swappable: a thin model boundary, an eval harness that measures agent work, and guardrails that don't move.

Jun 10, 2026 Read article

AI Engineering9 min read

Design is a model capability now - and why that changes agent UX

Frontier models have real design taste. Constrain it to your design tokens and components and the model becomes a fast implementer of your system - on-brand by construction.

Jun 24, 2026 Read article

Protocols11 min read

A2A in production - tasks, push, and security beyond the demo

The A2A features a hello-world skips but a real multi-agent system can't: push notifications, resubscription, security schemes on the agent card, and typed extensions.

Jul 15, 2026 Read article

Security12 min read

Letting an agent act as you - safely

Autonomy without guardrails is a liability. The three decisions that make agent actions shippable: forwarded identity, human-in-the-loop on writes, and an immutable audit trail - at scale.

Aug 5, 2026 Read article