Grupo Boticário

Specialist II, AI Engineer – Tech Demand and Operations

Grupo Boticário

full-time

Posted on:

Location Type: Remote

Location: Brasil

Visit company website

Explore more

✨ AI Apply
Apply

Tech Stack

ApacheAWSCloudGoogle Cloud PlatformJavaKotlinPython

About the role

  • The Demand Planning team at Grupo Boticário seeks to optimize production accuracy through a multidisciplinary structure that evolves from statistical models to an autonomous platform where AI acts as a strategic copilot for the business. The main challenge with Generative AI is ensuring technical accuracy and mitigating 'hallucinations' in a context that requires mathematical rigor so the AI can effectively support guidance and decision-making.
  • Design and implement Multi-Cloud AI Agent architectures: Be responsible for the technical design and implementation of autonomous agents operating within our Modern Hybrid Mesh architecture. You will orchestrate fast-inference models (such as Claude 3.5 Sonnet and DeepSeek-R1) via Amazon Bedrock for real-time on-screen interactions, and deep-reasoning models (such as Gemini 3 Pro) on GCP Vertex AI for batch analyses with context windows up to 1 million tokens.
  • Context Engineering and Advanced RAG: Design and optimize Retrieval-Augmented Generation (RAG) pipelines using Amazon Bedrock Knowledge Bases. Implement intelligent extraction techniques and dynamic metadata filtering to ensure agents accurately retrieve company inventory policies and lead times.
  • Develop Agents' reasoning capabilities: Build interactive copilots for demand planning using the Thought Signatures pattern to preserve AI reasoning cohesion in complex flows, ensuring deterministic and auditable responses based on structured data (Iceberg/RDS).
  • Asynchronous Integration and Orchestration: Develop secure communication between AI agents and our Kotlin microservices. For long-running tasks, you will integrate agentive flows with AWS Step Functions using the Callback pattern, eliminating idle time and ensuring fault tolerance.
  • Productionize AI solutions (LLMOps and Security): Ensure governance, inference cost monitoring, and application security. Implement Guardrails in Bedrock and configure cross-cloud access using Workload Identity Federation (WIF) for secure communication between AWS and GCP without static credential traffic.
  • Collaboration and technical development: Actively collaborate with the Modeling Squad and the Application Squad, translating outputs from predictive statistical models (ARIMA/XGBoost) into textual insights and automated action suggestions for Supply Chain users.

Requirements

  • Strong experience in Generative AI: Hands-on, proven experience in architecting, deploying, and productionizing complex applications using leading LLMs in mission-critical environments.
  • Deep knowledge of Python and Agent frameworks: Advanced fluency in Python and mastery of agent-building ecosystems and LLM orchestration frameworks (e.g., LangChain, LlamaIndex, Amazon Bedrock AgentCore, Vertex AI Agent Builder).
  • Multi-Cloud AI experience: Experience with at least one major managed AI provider (AWS Bedrock or Google Vertex AI) and a deep understanding of the differences between real-time inference and batch processing.
  • RAG architecture and Vector Database expertise: Solid experience in semantic chunking, embedding models (e.g., Titan Multimodal) and optimizing search in vector databases (e.g., OpenSearch, pgvector, Pinecone).
  • Technical judgment and critical thinking: Ability to make autonomous architectural decisions (e.g., when to use fine-tuning vs. RAG vs. advanced Prompt Engineering) aimed at optimizing latency and cost per token.
  • Plus: Practical experience implementing Workload Identity Federation (WIF) between clouds.
  • Prior knowledge of the AWS Serverless ecosystem (Step Functions, Lambda, EventBridge).
  • Understanding of Supply Chain business rules, Demand Planning (S&OP/IBP), or Finance.
  • Experience with Google BigLake and federated access to data in Apache Iceberg format.
  • Familiarity with Kotlin or Java to facilitate interfacing with the Backend team (Rules Engine).
Benefits
  • Here, your health is a priority
  • Medical and dental plans
  • Medication assistance
  • Health allowance for family members
  • Free psychotherapy sessions
  • Telemedicine and second medical opinions
  • Free flu vaccine
  • Health care programs
  • Meals and nutrition
  • Meal voucher or local restaurant allowance (depending on work model)
  • Food allowance
  • Christmas food allowance
  • Ensuring well-being and quality of life in all aspects
  • Gym and fitness studio plan
  • Home office allowance (for hybrid and remote work models)
  • Pet health plan
  • Birthday day off
  • Up to 40% discount on our products
  • Employee benefits program (partnerships and discounts, multi-brand store, gym passes, courses and more)
  • Travel and accommodation programs
  • Family support
  • Childcare assistance
  • Child nutrition credit
  • Babysitter allowance
  • School supplies allowance
  • Legal, psychological and social counseling
  • Support for parents of children with special needs
  • Extended parental leave (180 days for mothers and 120 days for fathers)
  • Commuting support for on-site work
  • Transit voucher and parking (for hybrid and on-site models)
  • Financial protection and peace of mind
  • Life insurance
  • Bereavement support for loss of parents
  • Private pension plan
  • Payroll-deductible loans
Applicant Tracking System Keywords

Tip: use these terms in your resume and cover letter to boost ATS matches.

Hard Skills & Tools
Generative AIPythonAgent frameworksLLM orchestration frameworksRAG architectureVector Databasesemantic chunkingembedding modelsWorkload Identity FederationAWS Serverless ecosystem
Soft Skills
technical judgmentcritical thinkingcollaboration