Description

You architect high-reliability inference systems to solve hallucination problems in Large Language Models, building proprietary IP beyond off-the-shelf solutions.

Responsibilities

• Define the strategy for domain adaptation and long-context reasoning using first-principles analysis to select RAG, Fine-Tuning, or novel methods.
• Design and build proprietary evaluation frameworks to rigorously measure model performance and safety before client deployment.
• Mentor the engineering team on the mathematical underpinnings of Transformer architectures and current SOTA research.
• Build custom inference pipelines, optimize vector search algorithms, or design complex retrieval systems.

Required Skills

• 8+ years of experience deploying complex ML research into production environments.
• Deep understanding of LLM failure modes: long-context recall, reasoning drift, and hallucination triggers.
• Proficiency in Python, writing clean, modular, object-oriented code.
• Experience building custom inference pipelines and designing complex retrieval systems.
• Familiarity with LLM Orchestration frameworks (LangChain, DSPy, LlamaIndex).
• Ability to mathematically constrain model outputs for high-fidelity results.
• Background in NLP, with experience applying RAG patterns.

Preferred Skills

• Interest in Code Generation, Program Analysis, or Semantic Parsing.
• Published research or technical blog posts on Applied NLP.

Key Skills

Education

Any Graduate