Knowledge Systems Architect

Meet Upside: We created Upside to transform brick-and-mortar commerce. Our technology uses the sophistication of online retail—profit measurement, attribution, and incrementality—to provide users with more value on their everyday purchases and brick-and-mortar businesses with new, profitable customers. We’ve helped millions of users earn 2 to 3 times more cashback than any other product, and hundreds of thousands of brick-and-mortar businesses earn measurable profit. Billions of dollars in commerce run through the Upside platform every year, and that value goes directly back to our retailer partners, the consumers they serve, and important sustainability initiatives. The Opportunity AI adoption doesn't fail because companies lack good tools. It fails because the organization isn't legible enough to use them. For AI to participate in real work drafting documentation, surfacing answers, flagging anomalies, executing workflows it needs material it can actually trust: artifacts that exist, are structured, are attributed to owners, and are fresh enough to act on. Most organizations skip this layer. They deploy AI on top of a knowledge environment full of stale pages, shadow Google Docs, unattributed decisions, and content no one quite believes and then wonder why the outputs can't be relied on. Upside is building differently. We're investing in the infrastructure layer that makes AI adoption compound rather than stall. The Knowledge Systems Architect owns that layer. This isn't a writing role. It's a systems design role. The person we're looking for doesn't create content they build the conditions under which content creates itself, gets maintained automatically, and becomes more trustworthy over time. They make the organization legible to machines and to itself. The scope starts with an R&D focus and expands from there. Why This Role Exists Now Upside has strong documentation instincts in some teams and gaps in others. We have powerful tools Glean, Confluence, AI documentation agents but adoption is uneven and the workflows that would make them self-sustaining don't exist yet. Documentation still depends too much on heroic individual efforts. The Knowledge Systems Architect changes that. Instead of being the person who writes the thing or answers the Slack message, they're the person who designs the system so neither of those is necessary. You won’t be starting from scratch but from a partially-built foundation: some islands of good practice, some legacy sprawl, and AI capabilities that are ahead of our governance. A significant part of the job is turning that foundation into a coherent, durable system. This is a high-leverage, high-visibility role inside the R&D Intelligence, Systems and Enablement (RISE) team. You'll work directly with the VP of RISE and alongside Engineering, Product, and IO to make Upside's knowledge infrastructure a genuine competitive advantage. Key Initiatives 1. BUILD THE ARTIFACT LEGIBILITY FOUNDATION FOR R&D The hardest part of AI enablement isn't deploying agents it's ensuring the knowledge artifacts those agents rely on are actually trustworthy. That means documents that are structured, attributed, current, and verifiable: not pages nobody has touched in two years, decisions buried in Slack threads, or content people cite without quite believing. You'll map the current state of artifact legibility across R&D, identify the highest-leverage gaps, and design the interventions governance frameworks, lifecycle rules, ownership models, structural standards, AI agents that close them. This is the core of the role. 2. AUTOMATE THE DOCUMENTATION LIFECYCLE Documentation shouldn't require someone to remember to do it. The goal is a world where documentation happens as a natural output of how work gets done triggered by product releases, embedded in team workflows, quality-checked automatically. You'll own the design and rollout of AI-assisted documentation pipelines: integrating agents into the Product Development Lifecycle and release workflows, setting up automated review triggers, defining when human oversight is mandatory and when it isn't. We have early infrastructure here (Glean agents, a product documentation agent, etc.). Your job is to operationalize it and make it irreversible. 3. SOLVE THE ROUTING PROBLEM Today, requests for documentation help, knowledge infrastructure, and tool guidance flow to a person. That doesn't scale and it creates a single point of failure. You'll replace that with a system: a well-governed intake pathway, enabled by an AI agent, and self-serve guidance that handles the majority of inbound without a human in the loop. The goal isn't efficiency it's changing the mental model from "ask an expert" to "I do this well on my own." 4. BUILD KNOWLEDGE OBSERVABILITY You can't govern what you can't see. We need real-time visibility into the health of our knowledge environment: what's stale, what's trusted, what's being used, what's not, and which teams are owning…