As a Full-Stack Software Engineer at my clients AI Lab, you will build the core product — AI agents and automation pipelines that take real accounting work off practitioners' plates. This is a high-ownership, hands-on role where you work directly with the founders and PM, own features end-to-end from schema to deployment, and ship production-grade AI systems you can fully explain and defend. Responsibilities: Design and build AI agent workflows using LangChain and LangGraph for document-heavy accounting tasks like tax return processing and bookkeeping automation. Develop and maintain Django/DRF backend APIs and ReactJS frontend components across the full product stack. Build document processing pipelines — extraction, classification, and normalization of semi-structured financial documents (PDFs, W-2s, 1040s, K-1s). Implement human-in-the-loop review queues with confidence thresholds, fallback logic, and hard rule overrides for AI pipeline steps. Design async task workflows using Celery and Redis for long-running document jobs with retry logic, progress reporting, and failure visibility. Write tests for AI pipeline logic, document parsing edge cases, and multi-step agent flows where silent failures have real consequences. Collaborate directly with the PM to size work, surface technical constraints early, and challenge underspecified requirements before building.