Lead Software Engineer – AI Agent Systems

About the role

At Bumble we are building a world where relationships are healthy and equitable. The Bee AI team is creating an intelligent, AI‑powered product that works alongside Bumble Date to help members connect in more thoughtful, human ways while preserving the core magic of matching and conversation.

Key responsibilities

• Lead the design, build, and extension of production AI agent systems for the Bee product, focusing on reliability, extensibility, and maintainability.
• Architect and evolve Python‑based services and agent workflows using Pydantic, AI, and Google Cloud Platform.
• Define end‑to‑end agent pipelines, including prompt management, context handling, response schemas, fallback logic, and versioning.
• Establish robust evaluation approaches for agent quality, covering offline/online testing, experimentation frameworks, telemetry, and clear success criteria.
• Build repeatable monitoring, debugging, and observability patterns for agents in production.
• Partner with Product, Design, and Data teams to turn ambiguous opportunities into shipped features.

Required profile

• Typically 8–10 years of software engineering experience, with deep expertise in Python and building scalable backend services.
• Hands‑on experience with AI/ML‑powered product experiences, especially AI agents, prompt design, and evaluation strategies.
• Proven ability to design schema‑validated LLM interactions, manage prompts, and create deterministic fallbacks for production.
• Experience with cloud infrastructure on GCP and designing systems with strong observability and resilience.

Required skills

• Python
• Pydantic
• Google Cloud Platform (GCP)
• AI/ML, Large Language Models (LLM)
• Prompt engineering and management
• Schema validation
• Observability and monitoring
• Experimentation and evaluation frameworks