---
title: Luca Destefano Boer — AI Engineer & Senior Software Engineer (Backend, Applied AI, Cloud)
url: https://lucadestefanoboer.com/
language: en
alternate_pt: https://lucadestefanoboer.com/pt/index.md
updated: 2026-11-03
---

# Luca Destefano Boer

Senior Software Engineer and AI Engineer based in São Paulo, Brazil (UTC−3), working professionally in software engineering since January 2023. Focused on backend systems, applied AI (LLM agents, RAG pipelines and real-time voice agents) and cloud infrastructure on AWS and Kubernetes. Works in Python, Go and TypeScript. Open to remote Senior Backend and AI Engineer roles, full-time or contract, with availability for US business hours.

- Website: https://lucadestefanoboer.com/ (Portuguese: https://lucadestefanoboer.com/pt)
- Email: lucadestefanoboer@gmail.com
- LinkedIn: https://www.linkedin.com/in/luca-destefano-boer/
- GitHub: https://github.com/lucadboer

## About

Luca started as a mobile developer in 2023, building React Native apps that supported social aid initiatives, such as food basket distribution for families in need. He then moved to full-stack work on an AI assistants platform, shipping Node.js and TypeScript APIs, streaming chat interfaces and document RAG pipelines, with performance-critical parts in Go and Rust.

Today he is a Senior Software Engineer and Tech Lead of an AI agents platform at a US company that builds AI agents and automation workflows for clients across industries, from healthcare and education to high-volume invoicing. He owns systems end to end, from architecture and code reviews to evaluations, guardrails and observability, and works daily with AI-assisted development tools such as Claude Code and Codex.

## Expertise

### Backend engineering
- Python, Go, TypeScript
- Node.js and NestJS
- REST, gRPC and OpenAPI
- PostgreSQL, MongoDB, Redis
- Temporal workflows
- Distributed systems
- Clean Architecture and DDD
- Unit, integration and E2E tests

### Applied AI
- AI agents and LLM orchestration
- Tool and function calling
- RAG, embeddings, vector search
- MCP (Model Context Protocol) servers
- Real-time voice agents (LiveKit, Twilio SIP)
- Golden-dataset evaluations
- Prompt-injection guardrails
- Langfuse, Portkey, LiteLLM, OpenRouter
- OpenAI and Anthropic APIs, LangChain, LlamaIndex

### Cloud, serverless and infrastructure
- AWS Lambda (serverless), EC2, EKS
- SQS and S3
- Cognito, KMS, IAM
- Kubernetes and Docker
- Terraform
- CI/CD with GitHub Actions
- OpenTelemetry
- Grafana, Prometheus, Datadog

## AI engineering in production

What it takes to run LLM features in real products, beyond calling a model API. These are the parts Luca has built and operated himself.

### Retrieval and RAG
- RAG pipelines, from ingestion and chunking to retrieval
- Hybrid search (BM25 + dense vectors)
- Rerankers to sharpen retrieval
- Query rewriting and multi-query retrieval
- Semantic caching

### Agents and reliability
- Tool-using agents with gated, permissioned tools
- Multi-agent orchestration with specialized agents
- MCP servers
- Structured JSON outputs
- Long-term memory for assistants

### Evaluation and safety
- Golden-dataset eval suites that gate releases
- LLM-as-a-judge evaluations
- Prompt-injection guardrails, including voice input
- Privacy-safe tracing

### Cost, latency and operations
- Model routing through AI gateways (Portkey, LiteLLM)
- Prompt caching and spend caps
- Token streaming over SSE with reconnection handling
- Tracing every agent step (Langfuse, OpenTelemetry)
- Prompt versioning and rollback
- Local open models with Ollama (Qwen)

## Experience

### Future Secure AI — Senior Software Engineer, Tech Lead
Oct 2024 – Present · Remote from Brazil · https://www.futuresecure.ai

US AI company building AI agents and automation workflows for clients across industries, including healthcare, education and high-volume invoicing. Leads the internal AI agents platform since April 2026.
Stack: Python, Go, TypeScript, AWS, Kubernetes, LiveKit, Temporal, MCP.

### AIZZY — Fullstack Developer
Jun 2023 – Oct 2024 · Remote from Brazil · https://aizzy.ai

AI assistants platform. Node.js and TypeScript APIs, React streaming chat, document RAG pipelines, Go and Rust components and AWS Lambda workers.
Stack: TypeScript, Node.js, React, Next.js, Go, Rust, Python, AWS Lambda.

### Maker Experts — Mobile Developer
Jan 2023 – Jun 2023 · São Paulo, Brazil

React Native and TypeScript apps that supported social aid initiatives, such as food basket distribution, with offline-first storage, REST integrations and App Store and Play Store releases.
Stack: React Native, TypeScript, REST.

## Side project

### recepia — WhatsApp booking assistant (2026 – Present)
An assistant that books routine dental appointments over WhatsApp and hands anything unusual to reception. It is not live with a clinic yet, but it is built like a production service: a deterministic engine owns every write, a Claude agent only proposes actions through gated tools, and an eval suite with prompt-injection attacks gates every push.
Stack: TypeScript, PostgreSQL, Claude API, OpenTelemetry.
Source: https://github.com/lucadboer/recepia

## Education

- MBA, Software Architecture — Full Cycle — 2026 to 2027 (in progress)
- Postgraduate Specialization, Advanced Software Development with Golang — Full Cycle — 2023 to 2024
- Technology Degree (Tecnólogo), Systems Analysis and Development — Universidade Paulista (UNIP) — 2021 to 2023
- Fullstack Development, intensive course — Rocketseat — 2023

## Languages

- Portuguese: native
- English: advanced (B2/C1)