Senior AI Systems Engineer

• Convert recurring engineering and product work into pipelines that agents carry from trigger to resolution, so the team stops doing it by hand
• Connect production signals to action: when errors spike (picture a burst of 500s inside half a minute), your system diagnoses and remediates before anyone gets paged
• Lift agent output by giving them richer context, sharper tools and sturdier infrastructure
• Keep services and message flows dependable as traffic and customer count grow
• Shape what gets built: flag what doesn’t add up, pitch ideas drawn from customers and competitors, and prototype before anyone requests it
• Make your work visible to leadership while it’s happening, surfacing blockers early rather than at the finish line
• Scout new models and techniques and adopt the ones that earn their place

• 5+ years writing software professionally, including a recent stint at an early stage startup where results were yours to own
• Some time at a larger organization, so you can reason about systems at scale
• LLM and agent systems you shipped to real users, beyond demos
• AI writes the bulk of your code today; your energy goes into the architecture and automation around it
• You generate ideas unprompted and back them with product and business reasoning
• You narrate progress as you work and ask for help early
• Fluent English, spoken and written, at the level of daily conversations with founders

• Production experience with queues, pub/sub or other event based messaging
• AI driven test automation, internal developer tooling or self healing systems you built
• A public trail of your work: technical writing, side projects, open source or community threads
• Founding or early engineer experience at a VC backed US startup

Python, TypeScript, AWS, frontier LLMs and agent frameworks.

• Location: Brazil, São Paulo area preferred (hybrid office).
• Time Zone: BRT
• Salary: $70K to $75K/year USD
• Equity: Yes
• Type: Contract

A US based, venture backed B2B SaaS startup past its Series A, around 35 people, building an AI agent that learns unfamiliar enterprise apps and data on its own and carries out real business tasks inside them. Engineering is small with no management layers, and AI already writes most of the code.