ML Engineer

IBM Automation and AI group is seeking an experienced AI/ML Engineer specializing in Large Language Models (LLMs) to join the AI/ML Center of Excellence team. The role involves collaboration with cross-functional product teams to implement AI capabilities across IBM's Automation product portfolio.

• Demonstrated experience with NLP and large language models (e.g., transformer architectures) including model evaluation and algorithm design.
• Strong programming skills in Python and familiarity with ML frameworks (PyTorch, TensorFlow, or JAX).
• Experience with data processing pipelines and working with large datasets.
• Knowledge of MLOps practices and tools for model deployment and monitoring.
• Ability to work independently and collaborate across diverse teams.
• Experience with fine-tuning and prompt engineering for LLMs.
• Deep understanding of transformer architectures and attention mechanisms.
• Proficiency in vector databases and embedding technologies.
• Knowledge of model serving frameworks (TensorRT, ONNX, TorchServe).
• Familiarity with cloud platforms (IBM Cloud, AWS, Azure, GCP).

Data Collection and Management for LLM Evaluation and Training:

• Design and implement robust data collection pipelines for diverse LLM training datasets leveraging the IBM AI Model & Data Catalog.
• Develop data quality assessment frameworks to ensure training data meets IBM's high standards.
• Create annotation guidelines and workflows for specialized domain-specific datasets.
• Implement data governance protocols that ensure compliance with privacy regulations and ethical AI principles following the IBM Data & Model Governance process and tooling.
• Establish evaluation datasets and benchmarks to measure LLM performance across various use cases leveraging FM-Eval and Unitxt.

LLM Integration and Implementation:

• Architect solutions to integrate LLMs with IBM's existing and emerging products and ecosystem.
• Develop APIs and interfaces that enable seamless interaction between LLMs and other software components.
• Optimize LLM deployment for various computing environments (cloud, edge, on-premises).
• Implement techniques for model compression, quantization, and optimization to improve inference efficiency and minimize resource requirements.
• Design and implement prompt engineering frameworks for consistent LLM behavior across products.

AI/ML Best Practices and Innovation:

• Establish technical standards and best practices for AI/ML feature implementation.
• Create reusable components and design patterns for common LLM use cases.
• Develop monitoring systems to track model performance, drift, and potential biases.
• Research and implement techniques for responsible AI, including explainability and fairness.
• Collaborate with product teams to identify opportunities for AI-driven innovation.