Senior GenAI Engineer – Advanced RAG

Founded in 1989, Millennium is a global alternative investment management firm. Millennium seeks to pursue a diverse array of investment strategies across industry sectors, asset classes and geographies. The firm’s primary investment areas are Fundamental Equity, Equity Arbitrage, Fixed Income, Commodities and Quantitative Strategies. We solve hard and interesting problems at the intersection of computer science, finance, and mathematics. We are focused on innovating and rapidly applying innovations to real world scenarios. This enables engineers to work on interesting problems, learn quickly and have deep impact to the firm and the business.

At Millennium, we are redefining how investment decisions are made. We don't just look at balance sheets; we harness the chaos of the real world. By analyzing vast amounts of unstructured data—from news briefings and earnings call audio to regulatory documents—we provide our Portfolio Managers (PMs) with the "informational edge" (Alpha) they need to outperform the market.

As a Senior GenAI Engineer at Millennium, you will build the "brain" of our investment platform. You will not just be "calling an API"; you will be architecting the entire lifecycle of unstructured data—from raw SEC filings and broker research to a sophisticated, Graph-powered Retrieval-Augmented Generation (RAG) system.

Your goal is to transform messy, complex financial documents and multi-media content into searchable knowledge enabling our Portfolio Managers to perform deep-dive analysis that traditional search cannot handle.

Key Responsibilities

• Data "AI-Readiness": Build pipelines to ingest and normalize complex documents (PDFs, Transcripts, Filings). You will implement advanced parsing logic to accurately extract tables, hierarchical headers, and embedded charts.
• Enrichments and Knowledge Graph Construction: Move beyond flat vector search by building GraphRAG systems and advanced annotations such topics, keywords, sentiment, etc. You will extract entities (Companies, People, Metrics) and relationships from text to build a dynamic Knowledge Graph that captures the nuance of the financial markets and its temporal aspects.
• Advanced RAG Orchestration: Implement state-of-the-art RAG techniques, including:

• Contextual Chunking: Semantic and agentic chunking strategies that preserve document context.
• Multi-Stage Retrieval: Hybrid search (Keyword + Vector) and re-ranking pipelines.
• Query Transformation: Implementing query expansion (Multi-query), decomposition, and rewriting to handle complex investment prompts.

• Graph-Vector Hybrid Systems: Leverage Graph-traversal (Cypher/Gremlin) combined with vector similarity to provide holistic context to the LLM.
• Evaluation & Observability: Build "RAG Evaluation" frameworks (e.g., Ragas, TruLens) to measure faithfulness, relevance, and hallucination rates in an investment-grade environment.

Required Technical Skills

• Programming: Mastery of Python (for AI/ML workflows) and Java (for high-throughput backend services)
• LLM Frameworks: Deep experience with LangChain, LlamaIndex, Haystack, etc
• Graph Technologies: Proficiency in Graph Databases (e.g., Neo4j, AWS Neptune, etc) and GraphRAG implementation patterns.
• Document Intelligence: Experience with OCR and parsing tools (e.g., Unstructured.io, LlamaParse, AWS Textract, or LayoutLM).
• Vector Databases: Expertise in Pinecone, Milvus, Weaviate, Chroma, etc
• Pipeline Engineering: Experience building high-throughput data platforms (Kafka, Spark) to process millions of tokens in real-time.

Qualifications

• 5+ years of experience in Software Engineering, with at least 2 years focused on GenAI/NLP production systems.
• Proven track record of moving GenAI projects from "Notebook/PoC" to "Production Scale."
• Strong understanding of Embedding models, Context Windows, and Tokenization challenges.
• Experience in Financial Services or working with SEC/Financial documents is a significant plus.