SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Cobre is Latin America's leading instant B2B payments platform, solving the region's most complex money movement challenges through advanced financial infrastructure. The company enables instant business payments—local or international, direct or via API—from a single platform, serving fintechs, PSPs, banks, and finance teams.
You would join the AI Engineering team to build AI products for Internal Solutions, transforming Cobre's customer support into an agentic system. Currently, many client questions about integrations, API behavior, transaction status, and webhook payloads are resolved manually by support staff across various channels. Your role is to build an agentic layer that resolves what it can against Cobre's real systems and escalates cleanly to specialists when needed.
Key responsibilities include:
- Own the agentic support layer end-to-end, migrating email automation from ad hoc n8n flows to a single decision engine
- Build deterministic scaffolding around the nondeterministic AI core: event-driven pipelines, idempotency, retries, dead-letter queues, typed domain models, and audit trails
- Treat client-data isolation as a core design requirement from day one, ensuring no response exposes one client's data to another across any channel
- Integrate against Cobre's real sources of truth—documentation, public API, sandbox, Snowflake, and Salesforce—behind clean, well-tested interfaces
- Design escalation paths with confidence thresholds and handoff logic so specialists pick up at exactly the right moment
- Instrument systems before building: correlation IDs, dashboards for resolution rate and time-to-first-response, meaningful alerts
- Keep model and vendor choices behind an anti-corruption layer for flexibility
- Work AI-natively while maintaining normal engineering standards: AI assistants, MCP servers, and agents are daily tools but must be reviewed, tested, and owned like hand-written code
- Partner across Customer Success, Integrations, and AI Engineering teams
- Feed learnings back into the platform as reusable patterns and components
Required qualifications:
- 5+ years building and operating production backend services with real post-launch ownership
- Depth in at least one backend-focused language; willingness to work in Go (Cobre's stack is Go-first with Python)
- Distributed systems fundamentals: asynchronous and event-driven processing, at-least-once delivery, idempotency, eventual consistency
- Cloud-native AWS development: Lambda, DynamoDB, S3, SQS/SNS, containers, IaC, CI/CD
- Architectural judgment: hexagonal architecture and domain-driven design in practice
- Testing discipline: high coverage as a gate, table-driven and contract tests
