The opportunity
Affirm is reinventing credit to make it more honest and friendly, giving consumers the flexibility to buy now and pay later without any hidden fees or compounding interest.
What you'll do
Build and ship AI agents, APIs, and applications on Affirm's internal: platform (Snowpark Container Services / Quicksilver). You own the full lifecycle: architecture, containerization, networking, secrets, CI/CD, monitoring, and fixing what breaks.
Turn messy business requirements from People Operations stakeholders into: production systems. Integrate with Workday, Notion, and case management tools so AI surfaces real answers from governed content, not model guesses.
Navigate Affirm's existing security and data governance infrastructure to get: AI systems running safely on people data. RBACs, data classification, and access policies already exist, but connecting them across systems (Workday, Snowflake, case tools) is where it gets messy. You figure out what's allowed, build within those constraints, and make sure employee data stays where it's supposed to.
Design reliability infrastructure for multi-model LLM services. Structured: output validation, fallback chains, circuit breakers for external APIs, and quality controls that catch hallucination before users see it.
Work directly with non-technical stakeholders to scope problems, make: architecture decisions, and give honest assessments of what AI can and can't do. You translate in both directions.
Contribute to the team's shared Python codebase, dbt models, and Snowflake: infrastructure as part of a small, full-stack team that ships fast.
What they're looking for
- Own what you build. When something breaks in production, you diagnose and fix it.
- Software engineering foundation. You have built, deployed, and maintained: production applications. You understand version control (Git/GitHub), CI/CD pipelines, containerization, and what it takes to keep software running, not just written.
- Systems thinking and technical architecture. You understand how software: systems fit together: databases, APIs, authentication, hosting, deployment pipelines. You can make architecture decisions, evaluate trade-offs, and read code well enough to know when something is wrong. The team works primarily in Python, and you should be comfortable in it, but the ability to think in systems matters more than raw coding skill.
- Builder disposition. You have created something from nothing: a system, a tool, a platform, in an environment where nobody handed you a spec. You identified the problem, designed the solution, and shipped it.
