LLM Engineer

Mid-Level LLM Engineer

Job Description:
We are an innovative AI products company building next-generation intelligent systems for real-world applications. As we continue to grow our research capabilities, we are seeking a Mid-Level LLM Engineer with a strong interest in model analysis, interpretability, and knowledge extraction to join our technical team.
In this role, you will focus primarily on research-driven development around Large Language Models, working closely with senior engineers and researchers. Your work will emphasize understanding model behavior, improving transparency, and developing methods to extract and structure knowledge from complex, unstructured data, while contributing to production-aware solutions when needed.

Responsibilities:

• Advanced Model Interpretability & Analysis
• Research and apply techniques for advanced model interpretability, including analysis of internal representations, reasoning patterns, and failure modes in LLMs.
• Investigate how knowledge is encoded, retrieved, and transformed within Transformer-based models.
• Knowledge Extraction
• Design and experiment with methods to extract structured, high-quality knowledge from LLMs and unstructured data sources.
• Support improved retrieval and reasoning through interpretability-informed insights.
• Data Analysis & Preparation
• Analyze and preprocess transactional, user, and merchant data for use in LLM-based research pipelines.
• Develop effective serialization strategies for heterogeneous data types.
• Research-to-Production Collaboration.
• Collaborate with engineering teammates to translate research findings into practical system improvements.
• Contribute to experimentation, evaluation, and documentation of research outcomes.
• Model Optimization (Supportive Role)
• Assist with prompt engineering, fine-tuning experiments, and Retrieval-Augmented Generation (RAG) setups under guidance from senior team members.
• Help evaluate performance, robustness, and limitations of different approaches.

 
Required Skills & Qualifications:

• Solid understanding of Transformer architectures and modern LLM frameworks.
• Background in statistics, probability, and sampling methods.
• Python proficiency (comfortable writing research code and small production components).
• Experience working with machine learning experiments, evaluations, or research prototypes.
• Curiosity-driven mindset with the ability to reason about why models behave the way they do.
• Comfortable working independently on defined research tasks while collaborating within a small team.
• Clear communication skills for sharing research findings with both technical and non-technical stakeholders.
• Exposure to model interpretability, explainability, or analysis tooling.
• Familiarity with RAG pipelines or knowledge-intensive NLP tasks.