Founding Full Stack AI Engineer

About Propel Labs We’re Propel Labs, the new research and development arm of leading learning and development agency Creative Design Group. Our award-winning agency designs cutting-edge learning and development solutions for Fortune 500 companies. Join us as we look to build the future of learning and development solutions! We are building AI-native learning and development products for small and medium sized businesses. Beyond that, we keep the details close, and we will walk you through the products, the architecture, and the roadmap in detail once we are talking. What we can tell you now: there is a working product, real technical depth to the problem, and no engineering team yet. The Role You are the first engineering hire. You will write the majority of the code, make most of the architecture calls, and own the product end to end: data model, backend, AI pipeline, web app, mobile app, and deploys. This is a hands-on individual contributor role, not a management role. You will work directly with the Director of Product & AI Strategy on scope and priorities, and with our instructional design and creative teams on product functionality and features. As Propel Labs grows, there is a clear path into technical leadership, but that is earned by shipping value, not granted on day one. Be honest with yourself about whether you want this. Founding engineer work means ambiguity, unglamorous plumbing, and a lot of decisions made with incomplete information. It also means the architecture is yours. What You'll Build The AI layer: Generation and synthesis pipelines built on frontier models, retrieval over customer-supplied documents, and the evaluation harness that keeps output quality from regressing every time we change a prompt. This is applied LLM engineering. We build on foundation models; we do not train them.The product surface: A multi-tenant SaaS web application plus a mobile app, along with internal authoring tools and customer-facing admin and reporting views.The platform underneath: Data model and tenant isolation, authentication and role-based access, background job processing for long-running generation work, usage metering (AI costs scale with usage, so this matters commercially), billing, and CI/CD.Ingestion: Document parsing across PDF, DOCX, and slides, plus web research to enrich generated output. Both pull in untrusted content, so both need to be handled defensively. What We're Looking For Full-stack SaaS engineering 2-3 years shipping production web applications, including at least one multi-tenant SaaS product you helped architect rather than inherited.Strong TypeScript and Python. Strong React experience — complex state, data fetching and caching, and keeping a large single-page app fast and maintainable.Relational data modeling in Postgres: schema design, migrations, indexing, query performance. You can explain how you would enforce tenant isolation and why you chose that approach.Mobile delivery experience, ideally React Native or Expo, including the parts people underestimate (builds, store review, OTA updates, offline state).Async and background processing: queues, workers, webhooks, retries, idempotency. Generation jobs take minutes, not milliseconds.Cloud infrastructure and operations: containers, environments, secrets management, observability, CI/CD. You have been on call for something you built. Applied AI engineering You have built and shipped a production RAG system, not a demo. You can talk specifically about chunking strategy, embedding model choice, hybrid search versus pure vector search, reranking, and how you measured whether retrieval was actually working.Prompt and context engineering as an engineering discipline: structured outputs, tool and function calling, multi-step orchestration, managing context windows, and making non-deterministic output behave inside a system that has to be reliable.Evaluation. You have built golden datasets and regression tests for LLM output, and you have opinions about where LLM-as-judge is useful and where it lies to you.Cost and latency as first-class constraints: token budgeting, caching, streaming, model routing, knowing when the cheap model is sufficient.Security in AI systems, specifically prompt injection through parsed documents and scraped web content, plus PII handling in prompts and logs.Comfortable in Python for the AI and data layer. How you work You use AI coding tools aggressively and well. This is not a preference, it is how we operate. We expect you to ship at a pace that assumes them, and to know where they produce confidently wrong code that needs a human.You can explain a technical tradeoff to a non-technical stakeholder without condescension or hand-waving. You will be doing this constantly.You make good decisions with incomplete requirements, and you ask rather than guess when the stakes are high.You care about the outcome for the end user, not just closing the ticket.Degree in computer science or a related field, or equivalent practical experience. We care about what you have built. Nice to Have EdTech, HR tech, or LMS experience, including standards like SCORM or xAPIRow-level security policy design at scaleEvent pipelines and analytics: turning behavioral data into something a customer will act onEarly-stage or founding-engineer experience Your First Six Months Month 1: Get deep in the existing codebase and product. Deliver an honest technical assessment: what to keep, what to rebuild, what will break at scaleMonths 2-3: Own and ship a significant feature end to end. Establish the evaluation harness for AI output qualityMonths 4-6: Harden the platform for real customers. Multi-tenant security, usage metering, monitoring, and a deploy process you trust on a Friday Compensation and Benefits $100,000.00 - $120,000.00 annually, plus bonus and future equity participation opportunitiesHealth, dental, vision, life insurance and 401(k) benefitsYou are joining an established, profitable agency with a long client history, not a company with 14 months of runway, and you are building the thing meant to become its next act. How to Apply Send us: Your resumeA link to your GitHub, or to something you have shipped that we can actually look atA short note (a few paragraphs, no cover letter theater) covering one AI feature you built and put in front of real users: what it did, how you evaluated whether it worked, and what you would build differently now We read every one of these. The third item is what we read most closely. Our process: an intro conversation, a technical assessment, a technical conversation about your own work and our architecture, a short project, and a final conversation with the leadership team. We share full product and architecture detail under mutual NDA at the technical stage.