Agentic AI Systems Engineer / Architect

About this role We're building an agentic AI system that integrates deeply with external services - payments, document workflows,and communications - and operates as a realtime, cloud native platform. We'reless interested in years of experience checkboxes and more interested in people who have actually built and shipped systems like this, and who understand why the pieces fit together, not just how to wire them up. You should be comfortable explaining tradeoffs in your past work, not just listing technologies you've touched. What you'll be doing Building and operating the backend,integrations, and infrastructure for an agentic AI system - includingthird-party API integrations, realtime frontend communication, and AWS based infrastructure. What we're looking for We care about demonstrated ability,not credentials. For each area below, be ready to show us something real - arepo, a deployed project, a design doc, or a walkthrough of a system you built. 1. Applied AI / agentic systemsexperience You've built with LLMs and agent frameworks in a real project, with agents reasoning, planning, and taking multi-step actions. You understand the practical challenges of designingpractical, non-deterministic systems for enterprises, including dealing with evaluations, handling failures or hallucinations, and when humans should be inthe loop. 2. Third-party API integration(backend) You've built and shipped integrations involving external, stateful services - e.g., payment processing, documentattestation/e-signature workflows, or transactional email and/orPlaywright-based outreach. We want to see how you handled failure modes, webhooks, idempotency, or retries - not just that you connected an SDK. 3. Real-time frontend communication You've implemented websockets or long polling in a real project andcan articulate the tradeoffs - when you'd choose one over the other, and what broke or what you learned. 4. AWS cloud infrastructure You have broad, hands-on experience across AWS - not limited to oneservice or pattern. This includes serverless/microservices work (Lambda,Athena, CloudWatch, S3, etc.), as well as traditional compute (EC2, Lightsail,etc.), and the underlying account/network layer (IAM roles and policies, VPCdesign, security groups). We're looking for someone who has architected systems that span multiple AWS services and made deliberate choicesabout which service fits which job - not someone who's only worked within asingle pattern. 5. Backend service development(FastAPI) You've built and deployed a non-trivialFastAPI service - async request handling, auth, background/task processing - ina real or production adjacent context. Experience configuring API Gateways is aplus, but on its own doesn't substitute for having built the underlyingservice. 6. Systems thinking andcommunication You can explain not just what you built, but why - the tradeoffs, constraints, and intent behind your design decisions. Just as importantly, you can adapt that explanation to your audience: translating technical decisions into business impact for non-technicalstakeholders, and going deep on architecture and tradeoffs with technicalpeers. We'll ask you to walk us through a past project twice - once as ifbriefing a business stakeholder, and once as if briefing a senior engineer. What to send us Instead of a traditional application,please include: ● Links to relevant repos, deployed projects, or case studies ● A short note on one project where you'd say the why mattered as much as the how