Ale Zanello

I'm Ale Zanello (Alejandro), an AI Engineer based in Córdoba, Argentina, working remotely worldwide. I have 6+ years of experience shipping production software across mobile, backend and web, and I now focus on agent runtimes, Multi-Agent Systems systems and generative AI. I build LLM systems with LangChain, LangGraph, RAG (Retrieval-Augmented Generation) and the Model Context Protocol (MCP (Model Context Protocol)) in TypeScript and Python, and I'm the author of Akashi, an open-source agent framework for Dart and Flutter.

How my career evolved

My path into AI engineering runs through operations, not a straight line from a CS lab — which is why I care about systems that real teams can run.

1. Operations. I started as a Branch Office Manager at California Fitness in Cancún, running branch operations, staff training, inventory and provider relationships.
2. IT and CRM. I moved into the IT organization as IT Manager, building and maintaining a fully custom web CRM in PHP and MySQL, introducing QA automation with Selenium and Cypress, and replacing manual processes with automated workflows.
3. Low-code. At Detecta I joined as a LowCode Specialist, delivering production mobile apps with FlutterFlow and Flutter on Firebase.
4. Product engineering. As Software Engineer at Detecta, I built and shipped the Detecta Security Platform — a triple-platform product with real-time GPS fleet tracking and AWS microservices built with NestJS.
5. Agentic AI. Since 2024 I've worked at Somnio Software as FlutterFlow Developer / AI Native Engineer, and since 2025 as an independent engineer (Independent AI & Full-Stack Engineer (Freelance)) focused on agents, real-time systems and products.

What I build today

My freelance work spans five lines:

• Conversational AI Agents — multi-tenant agents with LangGraph checkpointing, an MCP (Model Context Protocol) server, Langfuse / OpenTelemetry tracing and automated Evals.
• AI Agent Products & Remote MCP — production agents on Google Vertex AI and LangGraph, served through a remote MCP (Model Context Protocol) server with OAuth 2.1.
• Real-time Data Systems — event-driven ingestion, streaming and replay pipelines in Python with a Rust low-latency component, plus applied ML on top: a fine-tuned time-series foundation model and policies trained with Reinforcement Learning.
• White-label Flutter Apps — 6+ apps shipped from shared codebases (see White-Labeling).
• SaaS Products for Small Businesses — Next.js + Supabase products for small businesses, including the public NovaMX CRM.

Open source

• Akashi — a provider-neutral agent framework for Dart and Flutter, built as a 14-package monorepo with 13 packages on pub.dev (160/160 pub points for the core package, 4 published provider adapters). It brings a streaming tool-calling runtime, multi-provider failover, Durable Execution with Human-in-the-Loop approvals, and OpenTelemetry tracing to the Dart ecosystem.
• Somnio AI Tools — core contributor to an open-source CLI that installs and runs AI agent skills, health checks and security audits across AI coding CLIs (15 merged PRs).
• FlutterFlow issue triage — core contributor to FlutterFlow's public issue triage (600+ issues).
• Labs that study one idea in depth: mcp-deep-dive, langgraph-langserve-starter, local-rag-chatbot and langgraph-ts-agents.
• Developer tooling: Stores Helper and Spec Dev Ralph.

How I work

I practice Spec-Driven Development: write the specification first, then execute it with AI agents such as Claude Code, so the agent has unambiguous context and less rework. I write about it on the blog — start with What Is Spec-Driven Development? A 2026 Primer and Spec-Driven Development with GSD: A Complete Guide, then the tool comparison in Spec-Driven Development: GSD vs Spec Kit vs OpenSpec. Recurring themes in my systems are Agent Orchestration, Tool Calling, Observability and Evals.

Credentials

• IBM Generative AI Engineering Professional Certificate — IBM Generative AI Engineering Professional Certificate (IBM via Coursera, October 2026).
• Agentic AI Engineering Specialization — Agentic AI Engineering Specialization (Edureka via Coursera, May 2026).
• AWS Certified Cloud Practitioner — AWS Certified Cloud Practitioner (May 2024).

My formal studies and languages are in Education & Languages: I speak Spanish natively and English at a full professional level.

Availability

I take freelance contracts for agentic AI, GenAI integration and full-stack product builds. Contact: [email protected] · GitHub · LinkedIn · Résumé.

Related

• Roles: Independent AI & Full-Stack Engineer (Freelance), Somnio Software, Detecta, California Fitness
• Core skills: LLMs, LangGraph, MCP (Model Context Protocol), RAG (Retrieval-Augmented Generation), Multi-Agent Systems, TypeScript, Python, Dart, Flutter, Next.js, Supabase
• Also: Reinforcement Learning, Fine-Tuning