Principal AI Engineer

Infogain; Plano (TX); On-site; USD 120,000 - 150,000; Full time

Job summary

A tech firm is seeking a Research-Grade Engineer with a Master's or PhD who can combine deep theoretical knowledge of NLP with the ability to architect scalable systems. The role involves designing high-reliability inference systems addressing complex challenges like long-context reasoning and hallucination in LLMs. Candidates must have 8+ years of practical experience, strong Python skills, and knowledge of advanced LLM internals. This role offers significant responsibility in advancing NLP methodologies and requires excellent problem-solving capabilities.

Qualifications

• 8+ years of experience applying ML research in production.
• Experience with custom inference pipelines.
• Understanding of LLM failure modes and constraining outputs.

Responsibilities

• Architect high-reliability inference systems to address hallucination in LLMs.
• Define NLP strategy for domain adaptation.
• Design evaluation frameworks for model performance.

Skills

Advanced degree in relevant field

Strong proficiency in Python

Experience with LLM internals

Ability to architect scalable systems

Education

Masters or PhD in Computer Science, AI, or related field

Tools

Python

Job description

We are seeking a Research-Grade Engineer (Masters/PhD preferred) who combines deep theoretical knowledge of NLP with the ability to architect scalable, production-ready systems.

You will design systems that can handle massive context windows, maintain semantic integrity across thousands of files, and deliver verifiable accuracy.

You will define the methodologies to constrain Generative AI with strict structural rules, answering the hard question: How do we build a system that possesses the flexibility of a neural network but the reliability of a compiler?

What You Will Do

• Architecture Design: Architect high-reliability inference systems that solve the "hallucination problem" inherent in Large Language Models. You will move beyond out-of-the-box solutions to build defensible, proprietary IP.
• Advanced NLP Strategy: Define the strategy for domain adaptation and long-context reasoning. You will perform first-principles analysis to select the right approach (RAG, Fine-Tuning, or novel methods) based on rigorous benchmarking.
• Evaluation & Verification: Design and build proprietary evaluation frameworks to rigorously measure the performance and safety of our models before they touch client code.
• Technical Standards: Mentor the engineering team on the mathematical underpinnings of Transformer architectures and current SOTA research.

What We Need

• Advanced Degree: Masters or PhD in Computer Science, AI, or related field (or equivalent top-tier research lab experience).
• Advanced LLM Internals: You understand the specific failure modes of modern architectures regarding long-context recall, reasoning drift, and hallucination triggers in complex logic. You don't just fine-tune; you know how to mathematically constrain model outputs to ensure high-fidelity results.
• Applied Research: 8+ years of experience, with a track record of taking complex ML research and deploying it into production environments.
• Beyond APIs: Experience building custom inference pipelines, optimizing vector search algorithms, or designing complex retrieval systems.
• Engineering Excellence: Strong proficiency in Python. You write clean, modular, object-oriented code, not just "notebook scripts."

Preferred Experience

• Interest in Code Generation, Program Analysis, or Semantic Parsing.
• Experience with open-source LLM orchestration (LangChain, DSPy, LlamaIndex) but with a critical understanding of their limitations.
• Published research or technical blog posts on Applied NLP.