System Architect

---

---

Objective To build a Retrieval-Augmented Generation (RAG) system that allows non-technical strategists (like me) to ask natural language questions about global music upload and streaming data from UnitedMasters, in order to identify cultural trends, emerging markets, and brand opportunities. This system should be intuitive, insight-driven, and aligned with real use cases in brand strategy, marketing, and cultural analysis. Desired Outcome - A structured, queryable system that enables me to: - Ask questions like “What markets are growing fastest for Nigerian music?” - See charts, lists, or simple narrative answers tied to real data - Compare markets based on track uploads, growth %, or listener patterns - Surface brand-relevant insights (e.g. tourism trends, creator hotspots, collaboration potential) - Easily update or expand data over time This will power an insights platform we're calling Pulse — used to help brands navigate culture and global creativity. The Data - We have access to internal UnitedMasters data representing: Track uploads (by date, city, country) - Streaming destinations (i.e. where music from X country is being played) - Artist-level and region-level growth over time - Raw exports that need to be cleaned, merged, and made analysis-ready We’ve already done early analysis in Python, but now want a scalable, interactive solution. What Needs to Be Built - Data model: Structure and clean the data (normalize countries, time frames, resolve duplications) - Indexing for RAG: Use vector databases or embeddings so the system can “understand” cultural questions - Prompt layer: Allow me to ask strategic questions (in plain English) and get structured answers - UI optional: Ideally, build an interface or notebook that lets me explore without code - Maintainability: System should be easy to update with new data or new question types Who We’re Looking For - Data scientist or ML engineer with experience in: - RAG pipelines (LangChain, LlamaIndex, etc.) - Pandas, vector databases (like Pinecone or FAISS) - Building usable, strategic insight tools from messy real-world datasets Bonus: Experience in music, culture, or marketing analytics Timeline - We’d like to scope and staff in the next 1–2 weeks - Ideal MVP system live in ~4–6 weeks

Keyword: Web Development

Price: $75.0

Amazon Web Services Web Development Python Solution Architecture Database Architecture API