Binaryx

The Project is unlocking capital with AI by transforming credit origination through an intelligent platform that enables lenders to automate manual bottlenecks, underwrite 5× more loans, and reduce risk.

Founded by a team with deep domain expertise and a strong execution track record, and backed by a successful pre-seed round and strategic partners, we are building foundational AI infrastructure for the $3T+ asset-based finance market.

We are seeking an experienced AI Engineer to build production-grade AI applications that can understand, analyze, and reason over complex financial data and documents. This role goes beyond orchestration and surface-level integration: you will work deeply with LLMs, NLP, document processing pipelines, retrieval systems, and evaluation frameworks to deliver reliable, interpretable credit insights.

Success in this role requires strong fundamentals in NLP, Transformer models, document understanding, and system evaluation, alongside hands-on experience building and operating LLM-powered systems in production.

Key Responsibilities

• Design and build LLM-based systems capable of multi-step analysis across structured data, unstructured documents, and external knowledge sources.
• Implement advanced analytical workflows with chain-of-thought, tree-of-thought, and hybrid reasoning approaches, to improve task accuracy and robustness.
• Build rigorous evaluation frameworks that go beyond simplistic benchmarks like LLM-as-a-judge.
• Architect memory, context management, and workflow orchestration layers to support complex pipelines with optional human-in-the-loop review.
• Design, implement, and evaluate RAG pipelines, including document chunking strategies, embeddings, retrievers, re-ranking, and semantic search.
• Develop context engineering techniques that systematically manage model inputs, structuring retrieved context, tool outputs, and memory to improve reliability and reduce hallucinations.
• Build and maintain document processing pipelines, including OCR integration, layout-aware parsing, table extraction, and normalization of financial and legal documents.
• Develop ETL and data engineering workflows spanning structured, unstructured, vector, and graph data.

• Deploy, monitor, and operate scalable AI services on AWS or Azure, integrating them into CI/CD and observability pipelines.

   

Qualifications

• 5+ years of professional experience in Python or JavaScript.
• Demonstrated experience building and deploying production-grade AI applications using LLMs.
• Strong fundamentals in NLP, including:
  • Text classification, information extraction, and semantic similarity
  • NLP evaluation metrics (precision, recall, F1, confusion matrices)
• Solid conceptual and practical understanding of Transformer architectures, including:
  • Tokenization and context windows
  • Attention mechanisms (at a high level)
  • Model limitations, failure modes, and trade-offs
• Hands-on experience with document processing systems, including PDFs, scanned documents, tables, and semi-structured data.
• Strong experience with semantic search, retrievers, and RAG architectures, including embedding strategies and vector databases.
• Familiarity with modern LLM frameworks and tooling (e.g., LangChain, LangGraph, or similar) and observability tools (LangSmith, Langfuse, Helicone).
• Strong data engineering foundation: ETL pipelines, SQL/NoSQL databases, vector stores, and graph data models.

• Experience deploying and maintaining AI services on AWS or Azure, with CI/CD, containerization, and infrastructure-as-code.

   

Nice to Have

• Experience with model fine-tuning or adaptation techniques (LoRA, QLoRA).
• Familiarity with knowledge graph construction and usage, including GraphRAG or graph-enhanced retrieval pipelines.
• Exposure to advanced workflow orchestration patterns and evaluation tooling.