AI Platform Engineer

🌍 Our Vision & Mission At Payflows, we’re building the AI-native operating system for finance teams. We connect banking, payments, AI agents, and procurement into one product so finance teams can: Stop copy-pasting between tools Trust their data Focus on decisions, not busywork We care about real usage, not slides. Everything we ship should make a controller’s or treasurer’s day meaningfully easier. 🤝 Why Join Us? We’re a small, experienced team who’ve earned our stripes at Airbnb, Luko, BNP Paribas, and Facebook. It’s not always easy—things move fast, and we expect a lot—but if you’re hungry for impact, you’ll find it here. 🛠️ We’re builders at heart. We move fast, think deeply, ship often — and we’re always hungry to learn more. You'll collaborate closely with product and tech to craft elegant backend systems that scale as fast as we do. 🏆 Why This Role Matters Our AI agents sit at the heart of the product and touch multiple domains: Intake & Orchestration Agent – Listens to users (Slack, Teams, email, product), gathers context & documents, and routes each request to the right agents / workflows. Accounting Review Agent – Extracts data from invoices, matches with POs, posts to ERPs, and flags anomalies. Support & Copilot Agent – Answers product and process questions, ingests docs for Q&A, does text-to-SQL for reporting and builds visuals. Banking Reconciliation Agent – Syncs bank transactions with accounting records and catches mismatches. Short-Term Liquidity Agent – Automates fund transfers to keep accounts topped up and optimally invested. The underlying problems are messy and complex, but the experience for the user should feel obvious and simple. As our AI Platform Engineer, you’ll be the force multiplier behind these agents: you’ll build the platform, tooling, and especially the intake/orchestration layer that makes it easy and safe to launch new agents, try new ideas, and learn from production. 🛠 What You’ll Do Industrialize Existing Agents Put in place monitoring, logging, and real-time feedback loops so we can see exactly how agents behave in the wild. Attack latency, reliability, and failure modes until the experience feels snappy and trustworthy. Help define and track quality metrics (success criteria, regressions, user feedback signals). Build the Intake & Agent Platform Design and implement the Intake & Orchestration layer: how user intents, context, and documents are captured, enriched, routed, and audited. Build frameworks and SDKs that let other Payflowers ship new agents quickly without reinventing infra each time. Create self-service tools for data ingestion, evaluation, rollback, and safe experimentation. Applied Research & ML in Production Explore new ideas in LLMs, tools, retrieval, and planning—with a bias toward what can realistically run in production. Turn scrappy prototypes into robust, observable services that can handle real customers and real money. Work with our data folks to build evaluation pipelines (offline and online) that keep us honest. Collaboration, Writing & Mentoring Partner with backend engineers, PMs, and designers to ship end-to-end experiences, not just models or APIs. Write clear RFCs and docs so others can build on your work. Help raise the bar on engineering quality through reviews, pairing, and sharing context, not just code. 🔎 What We’re Looking For Proven ML/AI in Production (2+ years exp) – You’ve shipped ML/AI or LLM-based features into production and learned from real-world usage and failures. Strong Coding & Architecture Skills (5+ years exp) – Comfortable in Python, with good instincts for designing microservices, APIs, and data/agent pipelines. Infrastructure Savvy – Experience with AWS, Docker, CI/CD. Bonus: serverless, event-driven systems, queues, and background workers. Obsessed with Observability & Quality – You naturally think about logs, traces, dashboards, and alerts, and you use them to continuously improve systems. Team Player, Low Ego – You do thorough reviews, ask good questions, and are happy to be challenged as long as it makes the product better. 🧩 Bonus points if you’ve worked on: B2B SaaS or finance/ERP tools BaaS APIs or payment infra Serverless, AWS, event-driven architecture exposure 💻 Our Stack Backend: Rust / Python / Flask / SQLAlchemy Frontend: Typescript / Vue3 / Amplify Data: PostgreSQL / DynamoDB / S3 Infra: AWS / Serverless / Docker / Airflow / Databricks Tooling: Git / GitLab / Sentry #J-18808-Ljbffr