Founding AI Engineer

## About the Role Build the intelligence layer behind creator commerce. Lumo is building the AI operating system for TikTok Shop and Instagram affiliate marketing. We help e-commerce brands discover creators, automate outreach, manage affiliate campaigns, and surface performance intelligence that currently lives across spreadsheets, dashboards, and manual agency workflows. Our vision is simple: replace repetitive creator operations with intelligent systems that continuously learn, improve, and scale. TikTok is already live with paying brands. Instagram is now in active development. And the AI layer is becoming the core differentiator of the platform. We are a very small team moving extremely fast: - CTO - Full Stack Engineer - 2 Founders This is not an "AI integration" role. You will own the intelligence systems that power the company. The Role We're looking for an engineer who genuinely loves building AI systems in production. Not research prototypes. Not isolated notebooks. Not "AI strategy." Real systems. Used daily by real customers. Integrated deeply into a live product. You'll own everything AI at Lumo: - creator ranking systems - semantic search - RAG infrastructure - campaign intelligence - anomaly detection - recommendation systems - AI copilots - retrieval pipelines - future proprietary models You'll work directly with the founders and have genuine ownership over the systems you build. No layers of management. No endless sprint ceremonies. No "waiting for approval." You scope it. You build it. You ship it. ## What You'll Build First 90 Days You'll extend the existing FitScore engine across Instagram. This includes incorporating: - Reels interaction rate - hook retention signals - follower growth velocity - creator consistency scoring - cross-platform GMV intelligence - audience quality signals You'll tune ranking weights, improve matching quality, and build evaluation systems that continuously learn as new creator performance data flows into the platform. This is one of the core systems behind Lumo. Ongoing You'll own the LLM intelligence layer powering the platform chatbot. Today it already uses RAG across: - creator data - campaign data - shop metrics - outreach performance You'll improve: - retrieval quality - context construction - hallucination resistance - metric accuracy - ranking pipelines - multi-source retrieval orchestration - long-context performance The goal is not a novelty chatbot. The goal is an AI system brands genuinely rely on to make business decisions. Ongoing You'll expand the performance intelligence engine across Instagram and email infrastructure. This includes systems that detect: - GMV anomalies - creator inactivity - campaign pacing issues - DM expiry windows - catalogue rejection problems - cold email deliverability drops - attribution inconsistencies Every alert generates natural-language insight explaining: - what happened - why it matters - what action should be taken Post-MVP Once enough proprietary campaign data exists, you'll help transition Lumo from OpenAI-dependent intelligence into proprietary ML systems trained on our own creator-performance dataset. This is the long-term moat. Our belief is that the company that owns the creator-commerce intelligence layer will become extremely difficult to replicate. You'll help build that system. ## Our Stack The AI layer is embedded directly into the application stack. This is not a separate ML platform team. ## What We're Looking For You've already shipped AI systems into production environments. You understand: - RAG systems end-to-end - embedding pipelines - chunking strategies - retrieval optimization - re-ranking - prompt engineering - output parsing - vector search tradeoffs - evaluation systems - async inference pipelines You have strong opinions about: - where RAG breaks - why cosine similarity alone is not enough - retrieval quality vs latency - hallucination prevention - ranking accuracy - production reliability You're comfortable working in a full-stack TypeScript environment, not just isolated model infrastructure. You should enjoy: - moving quickly - owning systems independently - building practical AI products - solving messy real-world data problems - working close to product and users ## Bonus Points Experience with: - fine-tuning or instruction tuning - recommendation systems - anomaly detection - ranking systems - martech/adtech systems - creator economy tooling - event-driven infrastructure - SaaS startups - AI agents or autonomous workflows ## What You Get - Ownership over the company's core intelligence systems - Direct collaboration with the founders - Equity in a fast-growing category - A highly technical and ambitious environment - Real product impact from day one - Production systems used by paying brands - A structured, well-documented codebase - The freedom to build, experiment, and ship quickly This is the kind of role where your work directly shapes the future direction of the company. If you enjoy building ambitious AI systems that operate in the real world, we'd love to speak.