Senior AI Engineer
Summary
Senior AI Engineer for a global data and AI consultancy in Brazil (staffed via Aldelia), working remotely from anywhere in Brazil. The role leads production LLM applications end to end: agents and multi-step workflows, RAG pipelines, MCP servers, observability and evals, using Python, LangChain/LangGraph, vector databases and cloud platforms.
About Aldelia
Aldelia is a Franco-British multinational focused on HR solutions and outsourced services.
About Our Client
Our client is a global data and AI consultancy with a strong presence in Brazil, built on a culture of empathy. It believes a healthy workplace is one where every voice is heard, respected, and valued, and is committed to building a diverse and inclusive environment for people of all ethnicities, colors, genders, sexual orientations, gender identities, abilities, and religious beliefs. Candidates must be based in Brazil.
The Role
Our client is looking for a Senior AI Engineer to lead the design and delivery of AI applications in production. The professional will work with language and multimodal models (text, image, audio, and documents), such as GPT, Claude, Gemini, Llama, and Mistral, and will help define how AI connects with real systems, data, and processes.
This role is for someone who has already taken AI to production and knows the challenge goes beyond the prompt: reliability, observability, continuous evaluation, cost, security, and scale.
Key Responsibilities
- Lead the technical design and implementation of end-to-end AI applications, from conception to production deployment
- Design and build AI agents and multi-step workflows that perform real tasks, using LangGraph, LangChain, or other frameworks, and know when not to use a framework at all
- Develop solutions with multimodal models for use cases such as document data extraction, image analysis, and audio and video processing
- Design and optimize RAG pipelines: chunking strategies, embeddings, hybrid search, reranking, and retrieval quality evaluation
- Build and integrate MCP (Model Context Protocol) servers to expose internal tools, APIs, and data to models in a secure and controlled way
- Define and implement the observability strategy for AI applications: call tracing, latency, cost, and token usage monitoring, alerting, and failure analysis with tools such as Langfuse, LangSmith, Arize Phoenix, or OpenTelemetry
- Structure offline and online evaluation frameworks (evals) to measure quality, detect regressions, and support decisions on prompts, models, and architecture
- Ensure security and governance: guardrails, protection against prompt injection, handling of sensitive data, and compliance
- Mentor junior developers, conduct code reviews, and promote AI engineering best practices within the team
- Collaborate with product and business teams to translate real problems into viable AI solutions, with clear expectations about what the technology can and cannot deliver
Requirements & Qualifications
- Solid software development experience with Python, including engineering best practices (testing, typing, CI/CD, clean code)
- 3+ years of software development experience
- Proven experience taking LLM-based applications to production, not just prototypes
- Command of concepts such as agents, tool calling, RAG, structured outputs, and advanced prompt engineering
- Experience with frameworks such as LangChain, LangGraph, LlamaIndex, or equivalents, with a critical view of their trade-offs
- Experience with observability and evaluation of AI applications
- Experience with vector databases (pgvector, Pinecone, Qdrant, Weaviate, etc.) and relational databases
- Experience with REST APIs, Docker, and cloud deployment (AWS, GCP, or Azure)
- Ability to communicate technical decisions clearly to both technical and non-technical audiences
- Advanced or fluent Portuguese and English (required): both languages are essential for comfortable and clear communication in daily work and in an international environment
Nice to Have
- Knowledge of software and systems architecture: distributed systems, microservices, event-driven architecture, queues and messaging, and integration patterns
- Experience defining reference architectures for AI, such as LLM gateway layers, model routing, semantic caching, fallback, and cost control
- Experience with scalability, resilience, and high availability in production environments
- Knowledge of JavaScript/TypeScript (e.g., Vercel AI SDK, Next.js)
- Experience with fine-tuning, distillation, or hosting of open-source models (vLLM, Ollama, TGI, etc.)
- Experience with infrastructure as code (Terraform, Pulumi) and Kubernetes
- Open-source contributions or published technical content on AI
Benefits & Perks
- Meal and transportation allowance
- Flexible work setup, with the option to work from home or anywhere in Brazil
- Gym membership benefit
- Life, health, and dental insurance
- Regular in-person team gatherings, with workshops, lectures, training, and social activities
- Access to coworking spaces
- Semi-annual performance evaluations, with opportunities for promotion
Skills
- Agentic AI
- AI
- API
- AWS
- Azure
- CI/CD
- Cloud
- Distributed Systems
- Docker
- Embeddings
- Event Driven Architecture
- Fine Tuning
- GCP
- Infrastructure as Code
- JavaScript
- Kubernetes
- LangChain
- Langfuse
- LangGraph
- LangSmith
- LlamaIndex
- LLM
- MCP
- Microservices
- Mistral
- Next.js
- Observability
- Ollama
- OpenTelemetry
- pgvector
- Pinecone
- Prompt Engineering
- Pulumi
- Python
- Qdrant
- RAG
- RDBMS
- REST
- Terraform
- TypeScript
- Vector Databases
- Vercel
- vLLM
- Weaviate
Sign in to read and join the discussion.