GeekHunter Logo

Plans

Login

EN

EN

3M

Senior Data Scientist — Agentic AI

On-site

Sumaré, SP, Brazil

Salary Range

BRL

Full Time Employee

Experience Level

Senior

Requirements

Affirmative role
6+ years of experience in the career
English advanced
LLM
Python

Desired Skills

Power apps
Copilot
Amazon Bedrock
AWS (EC2, Lambda, SQS)
Amazon Web Services (AWS)
SAP
Salesforce
Power BI

Tasks and Responsibilities

You will design, build and run multi-agent solutions that act on 3M's enterprise processes — from framing the use case with the business to operating it in production on Amazon Bedrock AgentCore. You own quality, cost, safety and value, and you raise the bar for the engineers around you.

Collaborate with innovative

3Mers around the world Global Analytics & Insights (A&I), part of 3M Global Business Services, builds the automation, AI and analytics products that 3M's functions run on. This is an on-site role at our Sumaré site, from where we deliver for stakeholders in the United States, Brazil and Spanish-speaking Latin America: multi-agent AI solutions on AWS, business applications on Power Platform and Red Hat OpenShift on AWS (ROSA), and the integrations that connect them to SAP, Salesforce and legacy systems — under enterprise-grade, SOX-aligned change control.

The impact you will make in this role

  • Design, build and operate multi-agent systems — supervisor, swarm, graph and agents-as-tools — on Amazon Bedrock AgentCore, from use-case framing to production.
  • Give agents safe, auditable action on SAP, Salesforce and legacy systems through MCP tools, AgentCore Gateway and Policy, events and — where no API exists — RPA or Power Automate.
  • Own context engineering: retrieval, memory strategy, knowledge graphs and context-window budgets, with measured groundedness.
  • Own agents after go-live: evaluation harnesses (AgentCore Evaluations or equivalent), CI regression gates, tracing, cost and latency budgets, and rollback. Senior Data Scientist
  • Implement guardrails for prompt injection, PII and toxicity, least-privilege permissions and human-inthe-loop approvals; document solutions for governance review.
  • Choose classic ML, forecasting or optimisation when it beats an LLM on accuracy, cost or explainability.
  • Frame use cases with business stakeholders, quantify their value, and mentor engineers and data scientists through design and code reviews.

Required qualifications

  • About 6–9 years in data science or ML engineering, including 2+ years building LLM or agent systems that reached production.
  • Advanced Python: reusable packages, APIs, async, testing and code review.
  • Designed and shipped a multi-agent or tool-using agent system with one framework (Strands Agents, LangGraph, CrewAI or similar) on a managed platform — Amazon Bedrock/AgentCore preferred.
  • Production RAG or context pipelines with quality metrics, plus evaluation and observability of LLM systems.
  • Bachelor's or master's degree in Computer Science, Data Science, Mathematics, Statistics, Engineering or a related field.
  • Fluent or advanced English, used with global stakeholders.

If you meet most — not all — of the required qualifications, we still encourage you to apply. Years of experience are a guide to the expected level, not a strict cut-off. Strong evidence of equivalent impact counts.

Additional qualifications (nice to have)

  • Amazon Bedrock AgentCore in production, including Gateway, Identity, Policy and Evaluations.
  • MCP servers and A2A agent-to-agent collaboration.
  • Agents acting on SAP (incl. Integration Suite) or Salesforce (incl. Agentforce).
  • ROSA/OpenShift or Kubernetes, and data engineering at scale (SQL, ETL/ELT).
  • Combining agents with Power Platform, Copilot Studio, RPA or Amazon Quick Suite. A
  • WS Certified Generative AI Developer – Professional or AWS Certified Machine Learning Engineer – Associate.
  • Spanish is a plus.

How you work

  • Owns outcomes, not just tasks — and says early when something is at risk.
  • Explains technical trade-offs clearly to non-technical colleagues.
  • Uses AI tools daily and with judgement: reviews, tests and verifies what they produce.
  • Works well across cultures and time zones, in English, Portuguese or Spanish.
  • Is curious about automation, innovation and emerging technology — and learns fast.

Supporting Your Well-being

3M offers many programs to help you live your best life – both physically and financially. To ensure competitive pay and benefits, 3M regularly benchmarks with other companies that are comparable in size and scope.

How to apply

Applications are made on this page, directly with the company. You need a CV in PDF and an email address to confirm the application.

Share job:

How to apply

Applications are made on this page, directly with the company. You need a CV in PDF and an email address to confirm the application.

Share job: