Description

You will lead bias mitigation strategies for AI systems, ensuring compliance with government policies and Executive Orders. You will examine legacy models, develop mitigation approaches, and collaborate with customers to remediate or rebuild systems.

Responsibilities

• Assess legacy AI systems for bias and develop appropriate mitigation approaches.
• Compare legacy and bias-mitigated models for performance, accuracy, and business objective achievement.
• Refine models until bias mitigation is achieved and parity with legacy models is met.
• Document findings, publish code to repositories, and prepare standard operating procedures.
• Communicate the intent of AI government policies and Executive Orders to leadership.

Required Skills

• Minimum 8 years of Data Scientist experience.
• Bachelor's degree in Statistics, Applied Mathematics, Computer Science, or Information Science.
• Experience with Machine Learning (ML) and data quality assessment, governance, and protection.
• Knowledge of TensorFlow, PyTorch, and scikit-learn.
• Experience with Natural Language Processing (NLP), including spaCy, NLTK, BERT, and RoBERTa.
• Familiarity with ethical AI considerations including fairness, transparency, and accountability.
• Understanding of the full system development lifecycle and production-worthy AI systems.
• Proficiency with JavaScript and ES6+ features.
• Ability to obtain and maintain a Public Trust.

Preferred Skills

• Experience with IBM AI Fairness 360 (AIF360).
• Knowledge of Azure ML.

Key Skills

Education

Bachelor's Degree