Forward Deployed Engineer

The Role

You’ll operate at the frontier of our product — discovering what customers need, prototyping solutions, validating them in the field, and building them to production quality. Then you do it again.

The cycle:

• Discover — Work with customers to understand their GCP cost challenges, and bring your own ideas based on your engineering experience. Identify patterns, pain points, and opportunities that our platform doesn’t yet address.
• Prototype — Build working proof-of-concepts fast. Test assumptions with real data in real environments.
• Validate — Put prototypes in front of customers. Get honest feedback. Iterate or kill.
• Ship — Take validated solutions to production. Write the code, set up the infrastructure, make it reliable.
• Repeat — Move to the next opportunity. The cycle is the role.

Your initial focus will be BigQuery cost optimization — our primary optimization surface today. You’ll work deeply with BigQuery’s reservation model, slot management, job scheduling, and pricing mechanics. As the platform expands to cover more GCP services, so will your scope.

What You’ll Actually Do

• Research customer GCP environments to identify new optimization opportunities
• Build prototypes and proof-of-concepts to test optimization strategies
• Validate solutions directly with customers — run pilots, gather feedback, iterate
• Ship production-grade features — backend services, data pipelines, infrastructure-as-code
• Work with BigQuery APIs, INFORMATION_SCHEMA, reservation management, and job analysis at scale
• Write Terraform to manage customer-facing infrastructure on GCP
• Feed field insights back into product direction — you’ll have more customer context than anyone
• Represent Rabbit at industry events and conferences

Customer work is remote. You’ll communicate and collaborate with enterprise teams without needing to be on-site — but you’ll occasionally travel to industry events to represent Rabbit and connect with customers in person.

Must-have:

• 3+ years of software engineering experience — you build things, not just advise on them
• Deep GCP experience — you’ve operated production workloads on Google Cloud
• BigQuery expertise — you understand slots, reservations, pricing models, and query optimization
• Customer-facing instinct — you’re comfortable in enterprise conversations and can translate technical depth into business value
• Comfort with ambiguity — you’ll often define the problem before solving it
• Proficiency with agentic coding tools — we’re an AI-driven engineering team; you’re expected to leverage AI assistants and agentic workflows to move fast
• Autonomy — small team, no middle management, you own your work end-to-end

Strong plus:

• Experience with GCP cost management, billing APIs, or FinOps practices
• Familiarity with data pipeline tools (Airflow, dbt, Dataflow)
• Infrastructure-as-code experience (Terraform on GCP specifically)
• Background in data engineering or analytics engineering

Why This Role

• You decide what gets built next — not a PM, not a ticket queue
• Zero to production is your default mode, not maintaining existing systems
• Direct customer exposure means you see the real impact of what you ship
• Deep technical problems in cloud cost optimization — not CRUD apps
• Small team where there’s nowhere to hide and nothing is someone else’s job

How We Work — AI-First, Agentic by Default

Rabbit is an AI-first company. Agentic development practices are mandatory, not optional — Claude Code, Cursor, and agent-driven workflows are how we move. We measure output in shipped features, not commits.

Builders, not coders. If you’re attached to writing every line yourself, this won’t fit. If you orchestrate AI agents to compress idea-to-prototype from weeks to hours, you’ll thrive. The role is about judgment — what to build, for whom, and when it’s good enough — not implementation.

How to Apply

Send us your CV or LinkedIn profile, along with a brief note on something you built from scratch that reached production — what was the problem, what did you build, and what happened when real users hit it.