[ 02 / About ]
Antonio García — Software Architect & AI Solutions Builder.
I architect and scale distributed backends for companies across LATAM. High-throughput services in Go and .NET, asynchronous queues, and production LLM agents — wired together with strict SLOs and observability that holds under traffic spikes.
I don't just write code; I align system design with business metrics. The result: AI systems that ship to production, stay there, and pay for themselves.
/ 01 / 4
Designed for scale, not for CV-driven development. I map technical boundaries — DDD, event-driven, microservices — to real business needs.
/ How I build
/ 01 · System Architecture
Designed for scale, not for CV-driven development. I map technical boundaries — DDD, event-driven, microservices — to real business needs.
/ 02 · Core Backend & AI
High-throughput APIs in Go and .NET, wired to asynchronous queues and production LLM agents. The engines that move the data.
/ 03 · Cloud & Observability
Deploying to Kubernetes with strict SLOs. If it's in production, it's measured, monitored, and built to survive traffic spikes.
/ 04 · Client-side Engineering
Web and native interfaces built to the same rigor as the backend. Next.js on the web; Swift and Kotlin natively.
/ Certifications
[ 03 / Projects ]
LLM agents wired into business APIs
Architected RAG pipelines and lightweight LLM agents using Azure OpenAI. Automated unstructured data extraction, ticket triage, and contract analysis — directly impacting operational efficiency.
Real-time price intelligence for grocery shopping
Architected the core search and ingestion stack: vector embeddings for fuzzy product matching across supermarket catalogs, async ingestion pipelines normalizing thousands of SKUs, and attribute extraction from heterogeneous sources. Owned the native Android client and on-device barcode scanning end-to-end.
Distributed backends on managed Postgres
High-concurrency services built in Go and .NET over PostgreSQL, deployed to Azure or GCP. Engineered for strict performance budgets, seamless hand-offs, and rock-solid reliability.
Data extraction at scale
Reference implementation of a scraping pipeline using rotating proxies, retries and rate limiting. The patterns that make data collection survive production.
[ 04 / Experience ]
2019 — Now
@ atlantia.ai
Started as a .NET developer and grew into Technical Leadership. Evolved the stack: switched to Node.js for delivery velocity, then migrated to Go for memory efficiency under heavy loads. Architected the SRE practices, messaging queues, and currently lead end-to-end AI integrations via Azure OpenAI.
2016 — 2019
@ Centro Escolar Cedros
Built and shipped a school-management ERP from scratch — my first taste of leading a product end to end and learning what it really means to support real users.
2010 — 2016
@ SAP Business One · Brain Design · UNAM
Computer Engineering at UNAM FES Aragón. Apprenticeships in .NET / SAP B1 and freelance web work. The years that built the multi-stack reflexes.
[ 05 / Contact ]