Caio Cesar Dentello Perez Senior Artificial Intelligence Engineer Page 1
CAIO CESAR DENTELLO PEREZ
Senior Artificial Intelligence Engineer Agentic AI, RAG, LLM Integrations & Cloud Architecture Sao Paulo, SP, Brazil +55-11-920**-**** ************@*****.*** linkedin.com/in/caio-dentello-232506422 PROFESSIONAL SUMMARY
Senior artificial intelligence and full-stack engineer with 10+ years of experience building production software and AI- enabled systems for high-scale marketplace and SaaS environments. Hands-on with Python, LangGraph/LangChain, multi-agent orchestration, RAG, vector search, and OpenAI, Anthropic, and Gemini API integrations. Designed planner- executor-validator workflows, retrieval pipelines, tool integrations, evaluation gates, and cloud-native services across AWS and Azure. Combines agentic AI delivery with backend architecture, observability, security, and production reliability; a production automation platform reduced manual work by approximately 30%. CORE SKILLS
Agentic AI: LangGraph, LangChain, planner-executor-validator agents, tool calling, state and memory, human-in-the-loop workflows, structured outputs, prompt engineering, guardrails LLM APIs: OpenAI API, Anthropic Claude API, Gemini API, embeddings, streaming, function calling, provider abstraction, model fallback, prompt and model evaluation
RAG & Vector Data: document ingestion, chunking, embeddings, metadata filtering, semantic and hybrid search, reranking, citations, pgvector, Pinecone, Weaviate, Milvus
Cloud & Data: AWS, Azure Functions, Cosmos DB, PostgreSQL, MongoDB, Redis, Docker, Kubernetes, Terraform; Vertex AI and BigQuery at prototype/lab level
Protocols & Integration: REST APIs, event-driven services, secure tool APIs, MCP server/client concepts and prototype integrations; Google ADK at prototype/lab level
Engineering: Python, TypeScript, JavaScript, Node.js, React, Java, Spring Boot, Django, Ruby on Rails, SQL, CI/CD, observability, automated testing, secure delivery
PROFESSIONAL EXPERIENCE
iFood - Osasco, Sao Paulo, Brazil Mar 2021 - Present Senior Software Engineer AI, Agentic Systems & Backend Architecture
Lead architecture and delivery of AI-enabled marketplace and operational systems, combining Python services, LLM workflows, backend APIs, cloud infrastructure, and production quality controls.
Designed a production multi-agent platform with LangGraph/LangChain, RAG, vector search, tool APIs, and planner, executor, and validator agents to automate complex work while preserving auditability.
Integrated OpenAI, Anthropic, and Gemini models behind provider-aware interfaces with structured outputs, tool calling, retries, timeouts, fallbacks, logging, and cost and latency monitoring.
Built retrieval pipelines spanning ingestion, chunking, embeddings, metadata filters, semantic and hybrid search, reranking, source attribution, and regression evaluation for enterprise knowledge.
Introduced validation gates, confidence checks, human fallback, A/B prompt testing, tracing, and progressive rollout, reducing manual work by approximately 30% and improving consistency and reliability.
Mentor engineers through architecture reviews, debugging, incident analysis, secure delivery, and clear technical documentation across cross-functional product and platform teams. QuintoAndar - Sao Paulo, Brazil Mar 2019 - Feb 2021 Full Stack Software Engineer Cloud Services & Technical Quality
Architected customer-facing marketplace services using Python, Django, React, JavaScript, PostgreSQL, AWS, and REST APIs, balancing scalability, maintainability, reliability, and delivery speed.
Improved high-traffic APIs and data workflows through query analysis, caching, automated tests, observability, performance tuning, incident investigation, and root-cause analysis.
Reviewed implementation plans, pull requests, API contracts, and failure hypotheses, identifying missing edge cases, risky assumptions, and unnecessary complexity before production release. Agendor - Sao Paulo, Brazil Jan 2015 - Dec 2018
Software Engineer SaaS Backend & Integrations
Developed and modernized a B2B SaaS CRM using Ruby on Rails, JavaScript, PostgreSQL, and REST APIs, with emphasis on integration reliability, data modeling, and production support.
Refactored complex legacy components, optimized database queries, debugged production failures, and strengthened automated testing, security, readability, and long-term maintainability.
Contributed to architecture and code reviews, release decisions, backlog refinement, Agile delivery, mentoring, and technical documentation across product and engineering teams. SELECTED AI & CLOUD ENGINEERING PROJECTS
Production Multi-Agent RAG Platform
iFood Senior AI Engineer Jan 2023 - Present
Challenge: automate multi-step operational work without creating unsafe actions, unsupported answers, or opaque failure modes in business-critical workflows.
Solution: implemented a stateful planner-executor-validator architecture with retrieval, typed tools, validation checkpoints, human approval paths, retry policies, and durable execution state. Caio Cesar Dentello Perez Senior Artificial Intelligence Engineer Page 2
Quality: created curated eval datasets, trace-based debugging, regression suites, LLM-as-judge scoring, human review, and dashboards for task success, groundedness, latency, errors, and cost. LLM Provider, Tooling & Google Cloud Prototypes
Independent R&D Senior Engineer 2024 - Present
Production depth is strongest in LangGraph/LangChain, model API integrations, and vector RAG; prototype-level work covers Google ADK agent patterns, Gemini access through Vertex AI, and BigQuery-backed retrieval and analytics patterns.
Explored MCP server/client patterns for exposing tools and resources with typed schemas, authentication, timeouts, retries, input validation, idempotency, structured errors, and end-to-end tracing.
Compared provider routing, structured output behavior, tool-use accuracy, latency, and cost across OpenAI, Anthropic, and Gemini to support model selection and fallback decisions. Cloud Data & AI Backend Integration
Cross-project Senior Engineer 2021 - Present
Built cloud-native backend and integration services with Python and TypeScript, including Azure Functions, Cosmos DB, PostgreSQL, MongoDB, Redis, REST APIs, queues, and containerized deployment.
Designed partitioning and indexing strategies, idempotent operations, concurrency controls, observability, and secure service boundaries for scalable data-intensive workloads.
Automated delivery with Docker, Kubernetes, Terraform, CI/CD, monitoring, rollback planning, and production runbooks across AWS and Azure environments.
ROLE-ALIGNED TECHNOLOGY EXPERIENCE
LangChain / LangGraph - production: stateful graphs, agent routing, tool execution, durable workflow state, retries, validation nodes, and human approval paths
OpenAI / Anthropic / Gemini APIs - production: provider abstraction, structured outputs, embeddings, streaming, tool calling, fallback, observability, latency and cost controls Vector databases - production/project: pgvector, Pinecone, Weaviate, and Milvus with semantic and hybrid search, metadata filtering, reranking, and retrieval evaluation Cosmos DB - hands-on: Azure Functions integrations, partition and index design, idempotent writes, concurrency handling, query tuning, and operational monitoring
MCP - prototype/lab: server/client integration patterns, typed tools and resources, schema validation, authentication, timeouts, structured errors, and tracing
Google ADK / Vertex AI / BigQuery - prototype/lab: agent construction patterns, Gemini model access, and analytics or retrieval-backed workflow designs; production depth remains strongest in LangGraph/LangChain and AWS/Azure EDUCATION
FIAP - Faculdade de Informatica e Administracao Paulista Feb 2011 - Dec 2014 Bachelor's Degree in Information Systems Sao Paulo, Brazil