Staff Software Engineer

Pearson is seeking a Staff Software Engineer for its Pearson Software Group - Content Platform team. The role embeds with internal lines of business to discover problems, drive AI adoption, and build, deploy, and support production software end to end. The engineer will develop LLM-powered applications, iterate rapidly with stakeholders, and codify successful patterns for broader platform and engineering use.

Responsibilities:

• Going where the problem is. You will work directly with the internal stakeholders inside our different lines of business. Through real understanding of their pain points, their workflows, and the architecture of their current technical landscape, you'll design and build systems end to end for them. These are the people who will actually use what you build. You meet them on their ground, in their workflow, with their data
• Conducting real discovery. You're not waiting for a finished brief. You sit with users, watch them work, ask the questions, and translate what you hear into something buildable. The discovery is part of the engineering job, not a step that happens before it
• Building during the conversation. The first version often exists before the meeting ends. People react to working software very differently than they react to wireframes — your job is to give them something to react to as fast as possible
• Working from messy inputs. Spreadsheets, PDFs, screenshots, recorded calls, half-written documents, contradictory emails from two stakeholders. You read all of it and turn it into something coherent. You don't ask people to clean up their inputs before you start
• Shipping production AI applications. Beyond prototypes, you build real applications on top of frontier LLMs, agents, MCP servers, evaluation harnesses, retrieval systems, custom skills — that run in production against real workflows. You know how to evaluate what you've built, not just demo it
• Iterating fast across many engagements. A typical engagement runs through many versions in a short window. Requirements shift every cycle. You don't get attached. You ship the next one
• Owning the full stack. Frontend, backend, data, integrations, deployment, the URL you send the stakeholder, and the observability around it once it's live. There is usually no one else to hand pieces off to. If it isn't done, it's because you haven't done it yet
• Codifying what works. When you find a pattern that ships across engagements — a prompt structure, an evaluation harness, an integration shim, an agent template — you push it back to the platform and engineering teams so the next FDE doesn't rebuild it. You make the rest of us better, not just yourself faster
• Owning the relationship over time. The first prototype is the start, not the deliverable. You stay close to the line of business through the lifecycle of an engagement, identify new opportunities as they surface, and harden what graduates into production

Requirements:

• 8+ years of full-stack engineering experience, with meaningful time spent shipping production systems end to end. Background as a technical founder, FDE, or software engineer with consulting experience is all fair game
• Strong CS fundamentals. Data structures, algorithms, system design. You can whiteboard a service architecture, talk through tradeoffs, and not flinch at a coding interview
• Frontend depth. Modern React, TypeScript, real component architecture, state management you can defend, and the taste to build something that looks finished — not just functional
• Backend depth in order of importance (Node.JS, Java, and Python). API design, data modeling, auth, error handling. Comfortable owning a service from request handler to schema
• Production LLM experience. Advanced prompt engineering, agent development, evaluation frameworks, retrieval, and deployment at scale. You've shipped at least one real LLM-backed application that someone other than you used
• Fluency with AI-assisted development tools and agentic coding. This is not a side note. The work moves at a pace that assumes you're using these tools efficiently, accurately, and rapidly. You should be able to demonstrate a clear process that delivers measurable results
• Data fluency. Databases at a real working level, comfort with Python data tooling
• Cloud and deployment fluency. AWS, GCP, or Azure — enough to put a service on a real URL behind real auth without filing a ticket. CI/CD, containers, observability
• Integration experience. You've built things that talk to systems you didn't write. SSO, OAuth, third-party APIs, internal platforms, legacy databases
• High agency. You navigate ambiguity inside a complex organization without needing someone to clear your path. You make calls and own them
• High cooperation, low ego. Pearson is 30,000+ people. Forward deployed work crosses team lines constantly
• Communication skills that work in both directions. You can sit in a room of executives and engineers, hold both threads at once, and not lose either audience. You can speak to end users about their problem and translate that into something other engineering teams can actually build against
• Currency on what's changing. You stay close to the frontier of LLM capabilities, agent patterns, and AI product stacks
• Education: Bachelor's degree in Computer Science or equivalent combination of education, training, and professional experience