AI Builder

• EarnIn is making AI-native engineering a core capability — not an initiative, but how we design, build, and ship
• We’re not hiring a software engineer who dabbles in AI
• We’re hiring an AI builder who also writes great software — someone who looks at every step of how we design, build, test, and ship, and asks: why isn’t an agent doing this?
• This is not a side project
• The agents and harnesses you build here will run in production, making real decisions for real people — people who depend on EarnIn to access their pay when it matters most
• If that bar excites you rather than intimidates you, keep reading
• You’ll work at the intersection of platform engineering, developer experience, and applied AI — partnering with architects, domain leads, and product engineers to build the tools, patterns, and guardrails that make AI adoption fast, safe, and durable
• Agents that take real actions. You’ll design how agents think — prompts, reasoning chains, tool calls, and the full architecture beneath them. From MCP servers and agent scaffolding to context harnesses and a Skills Marketplace, you’re building the layer every squad at EarnIn builds on. Not a proof of concept. Production infrastructure
• A PDLC that doesn’t look like 2022. The product development lifecycle is overdue for a rethink. You’ll challenge each stage — from scoping and design through to review, testing, deployment, and monitoring — and replace manual friction with agentic workflows wherever it makes sense. Evaluation pipelines, automated PR hygiene, deployment gating, generation-to-merge metrics: you’ll build the scaffolding that makes governed AI fast, not slow
• Evals that actually mean something. You’ll own the evaluation infrastructure — building the pipelines, benchmarks, and quality gates that tell us whether our AI is working, degrading, or ready to ship. That means designing eval harnesses for AI-assisted workflows, setting generation-to-merge and review latency baselines, and making model quality visible and trustworthy across teams. If you’ve built evals that caught real problems before they hit production, you’ll fit right in
• Pilots that graduate to production. EarnIn’s engineering leads are running experiments. You’ll take what’s working and turn it into something the next team can fork and ship in a week. Less “interesting prototype,” more “thing we rely on.” You’ll build reusable libraries, templates, and reference implementations that give squads a running start on AI integration, so no one has to solve the same problem twice
• —-WHAT SUCCESS LOOKS LIKE—-
• Teams across engineering can ship AI-assisted features faster, with fewer rework loops
• The harnesses you build are actively used and well-documented
• AI pilot quality and safety metrics are visible, trustworthy, and improving

Benefits

• 401k
• FSA
• Mental health resources
• Employee Resource Groups
• Stipend for growth opportunities
• Pet friendly office
• Work from home opportunities
• Health insurance- Bachelor’s, Master’s, or PhD in Computer Science, Computer Engineering, or a related technical discipline, or equivalent industry experience. We care about what you’ve built, not where you studied
• 3+ years of full-time software engineering experience, with at least 2 years building tooling, platforms, or internal developer products
• Experience with MCP, LangChain, or comparable orchestration frameworks
• Strong opinions about developer experience and a track record of building things other engineers actually use
• Experience with open source LLM models
• Proficiency and comfort working across the stack when needed
• Hands-on experience with LLM integration patterns — prompt engineering, RAG pipelines, tool/function calling, and agent architectures
• Exposure to multi-step reasoning pipelines or human-in-the-loop system design
• Hands-on work with AI governance — bias evaluation, audit logging, model cards
• Experience in fintech or regulated/security-sensitive environments
• Hands-on experience with reinforcement learning — especially RLHF, RLAIF, or reward modeling in applied product contexts