Lead AI Engineer

We are building the "Self-Driving Car" for the music industry.

Label OS is pivoting from a service platform to an Agentic Automation Operating System. We are replacing the manual "Frankenstein Stack" of the music business (spreadsheets, emails, disjointed portals) with autonomous AI Agents that handle Scouting, Legal Administration, Royalty Finance, and Distribution.

We don't need chatbots. We need deterministic, reliable, enterprise-grade Agents that can be trusted with a record label's P&L.

The Role
As our Lead AI Engineer, you will own the architecture and execution of our "Agentic Core." You will move us from "Prototype" to "Production," solving the hardest problem in GenAI: Reliability.

You will be the architect behind our high-stakes agents (e.g., the Legal Admin Agent and Finance Agent), where "hallucinations" are not an option.

What You Will Build & Own

• The "Trust Engine" (Evals): You will build the automated evaluation pipelines (using frameworks like DeepEval or RAGAS) to test every agent before it ships. Your mandate is 99.9% reliability on structured tasks.
• Complex Orchestration: Design multi-step workflows (using LangGraph, DSPy, or custom state machines) where agents must plan, critique their own work, and execute tasks across multiple tools.
• COGS Optimization: You will own the "Unit Economics" of our AI. You will optimize context windows, prompt caching, and model routing to ensure we hit our SaaS margin targets.
• Structured Output: Ensure our agents don't just "chat," but output perfectly formatted JSON, XML, and legal syntax that integrates directly with our database and DSP supply chain. 

• 5+ years of software engineering experience, with at least 2+ years deeply focused on LLM/AI applications in production.
• Production RAG Experience: Proven track record of building and deploying Retrieval Augmented Generation systems that scale (beyond simple vector DB tutorials).
• Python Proficiency: Expert-level command of Python for backend development (FastAPI/Django/Flask).
• Orchestration Frameworks: Hands-on experience with tools like LangChain, LangGraph, DSPy, or similar custom implementations.
• Database Knowledge: Experience with Vector Databases (Pinecone, Milvus, or Postgres with pgvector) and relational database design. 

• The "Reliability Obsessive": You know that a prompt is not a product. You are obsessed with Evals, Unit Testing, and Guardrails.
• Engineer First: You build robust APIs, not just Jupyter notebooks.
• Pragmatic: You know when to use a massive model and when to use a Regex. You prioritize speed and cost-efficiency over "hype."

Why Join Now?

• Strategic Impact: You are not a cog in a machine. You are building the core IP that drives the platform's value.
• Greenfield Challenge: We have the data and the customers (via major strategic partnerships). You get to build the brain.
• Competitive Compensation: Attractive salary package reflecting the seniority of the role.