Senior Software Engineer (Backend)

Role Summary:

As a Senior Software Engineer, you’ll own major parts of our AI stack. You’ll prototype zero-to-one workflows, design multi-step and RAG-based pipelines, and push features into production that real attorneys use within days, not quarters. You’ll work across backend, orchestration, retrieval, and product-integrated AI systems.

Everything you ship will directly shape how legal teams review, revise, and negotiate contracts inside Word and the browser.

Responsibilities

• Own major product surfaces that rely on retrieval, structured reasoning, and multi-step AI pipelines
• Prototype, evaluate, and productionize new AI workflows for contract review, clause-level analysis, playbook enforcement, and multi-document tasks
• Improve our internal LLM orchestration layer, prompt-library, evaluation tooling, and RAG strategies
• Partner with founders and our in-house legal team to design and validate new features
• Build backend services that support multi-tenant enterprise customers and high-volume workloads
• Ship quickly: small PRs, fast iteration cycles, functional releases every week

Qualifications

• 7+ years of engineering experience
• Strong backend engineering skills and ability to build production systems end to end
• Experience shipping ML or LLM-powered products at scale
• Experience with retrieval, embeddings, experimentation, and evals
• Ability to design multi-step pipelines and agentic workflows
• Strong instincts around latency, reliability, and fallback strategies
• Ability to prototype in days and productionize in weeks
• High standards for correctness, reliability, and UX quality
• Bonus: Go, Next.js/React, Postgres/pgvector, Cloud Run, multi-provider LLM orchestration

What You'll Work On:

• Designing a multi-step agent that analyzes an entire contract, applies playbooks, generates redlines, and formats them into Word-ready output
• Improving a retrieval layer that handles 300+ page agreements or mixed corpuses of PDFs, templates, and prior negotiations
• Building a system that scores clause-level correctness using structured evaluation datasets
• Creating a streaming, long-running workflow for contract review with checkpoints, retries, and deterministic outputs
• Designing a prompt-library and golden-file test suite that guarantees consistent outputs across multiple LLM providers