AI AgentsVoice AIMemoryMultimodal AIContext EngineeringHuman-AI InteractionAgent Runtime
ENGINEER · AI SYSTEMS · BUILDER
I build intelligent systems.
Artificial Intelligence · Software Architecture · Intelligent Systems
I'm a software engineer focused on artificial intelligence, intelligent systems and software architecture. I explore how modern AI capabilities can be transformed into reliable, scalable products - from LLMs and agents to voice interfaces, memory, multimodal systems and distributed architectures.
Positioning
AI capabilities become valuable when they become systems.
My work sits between model behavior, product constraints and the architecture needed to make intelligent software dependable in the real world.
Selected Work
Projects exploring persistent context, agent runtimes, multimodal interaction and reusable AI application primitives.
NexaLab Agent SDK
Reusable primitives for capabilities, tools, memory, artifacts and agent execution.
Agent primitivesCapabilitiesTool executionAgent memoryArtifact generationExtensible architectureProvider-oriented design
What I'm exploring
A focused map of the questions and systems I keep returning to.
Artificial Intelligence
Understanding model capabilities, limits and product implications.
AI Engineering
From models and prototypes to reliable systems in production.
Intelligent Systems
Systems that perceive, reason, remember and act.
Software Architecture
Designing the technical shape that lets products scale and evolve.
Human-AI Interaction
Exploring more natural ways for humans and intelligent systems to interact.
Future of Computing
The shift from static software toward contextual, adaptive systems.
AI engineering beyond the API call
Useful AI products require orchestration, memory, evaluations, tool use, routing, observability and careful latency decisions. The model matters, but the surrounding system often determines whether the experience feels reliable.
LLMsAgentsMemoryRAGEmbeddingsContext EngineeringTool UseModel RoutingVoice AIMultimodal AIEvalsObservabilityLatencyReliabilityAI Infrastructure
Architecture for intelligent products
The same systems need strong engineering foundations: APIs, event-driven flows, real-time channels, databases, multitenancy, integrations and operational visibility.
Distributed systemsEvent-driven architecturesREST APIsSSEWebSocketsPostgreSQLSQL ServerMultitenancySaaS architecturesObservabilityScalabilityEnterprise integrationsDockerTypeScriptNode.js.NET
Latest Writing
Recent posts from the blog, plus planned essays on agents, memory, voice, context and the architecture of AI systems.
AI Agents Need State
A draft note on agents as stateful systems rather than stateless prompt wrappers.
RAG Is Not Memory
A draft note on why retrieval can support an AI system without becoming memory by itself.
The Architecture of a Persistent AI Agent
A draft outline for the components behind persistent context, tools, memory and execution.
Context Engineering Is Becoming Software Engineering
A draft note on context as a core system design concern.
Latency Is Part of the Intelligence
A draft note on why speed is not just performance but product behavior.
Long Context Doesn't Replace Memory
A draft note on the difference between larger windows and durable system memory.
What MCP Actually Solves
A draft note on protocol boundaries, tools and interoperability for AI systems.
Why Voice Changes AI
A draft note on latency, turn-taking and presence in voice-first AI interfaces.
About
I'm Abraham Huerta, a senior software engineer focused on artificial intelligence, intelligent systems and software architecture.
My work lives at the intersection of AI and software engineering. I'm interested not only in what modern models can do, but in the systems required to turn those capabilities into reliable products.
Read moreAI Systems
Software Architecture
Backend Engineering
Real-time Systems
Enterprise Integrations
Product Engineering
Let's build what's next.
For collaboration, technical conversations or relevant opportunities, email is the clearest path.