Software Engineer I, AI Specialist

Warner Bros. Discovery•Atlanta, GA

About The Position

The AI Enablement & Machine Learning team at CNN is accelerating our digital transformation through strategic applications of machine learning and AI technologies. The AI Systems & Enablement group within this team builds and operates AI-powered applications for CNN's audiences and internal teams, and drives AI adoption across the organization through tool governance, training, and enablement programs. Current products include article summaries, content and product classification, weather summaries, and brand safety classification. The team also leads CNN's AI enablement efforts — including AI Practices Guild leadership, Claude Code rollout, foundation model governance, and productivity measurement. Our vision is that CNN delivers AI-powered features that enhance user experience and internal productivity with systems that are reliable, accurate, and trustworthy — meeting the high standards our audiences expect from a news organization — while equipping the broader organization with the tools, governance, and skills to work effectively with AI.

Requirements

2–4 years of experience in content operations, content classification, editorial workflows, taxonomy management, or a related field
Demonstrated experience with prompt engineering for large language models in a professional or academic setting
Experience designing evaluation criteria or quality assessment workflows for content or AI outputs
Strong understanding of how LLMs work — including their capabilities, limitations, stochastic behavior, and cost implications
Excellent written and verbal communication skills, with the ability to explain technical AI concepts to non-technical audiences
Detail-oriented with strong analytical skills and comfort working with evaluation data and quality metrics
A collaborative mindset and comfort working closely with software engineers and editorial stakeholders
Comfort navigating ambiguity in a fast-evolving space — applied AI best practices are still being established

Nice To Haves

Degree or certification in AI, machine learning, data science, or a related field
Experience with content classification, taxonomy, or content management systems
Experience in a news, media, or content-heavy environment where accuracy and trust are paramount
Familiarity with AI evaluation frameworks, annotation tools, or human-in-the-loop methodologies
Experience working with cross-functional teams that include editorial and engineering stakeholders
Understanding of information retrieval concepts and how they apply to AI-powered features
Basic scripting skills (e.g., Python) for data analysis or automation of evaluation workflows
Experience with requirements gathering, process documentation, or business analysis in a technical environment

Responsibilities

Own the prompt engineering lifecycle for assigned AI features — from initial design through iteration, testing, and production maintenance
Build and maintain evaluation frameworks that measure AI output quality against CNN's editorial and accuracy standards
Design and run human review workflows, ensuring feedback is structured and actionable for engineering and ML partners
Partner closely with engineers on your team who handle code, architecture, caching, optimization, and deployment — providing the content expertise, evaluation data, and quality guidance that shapes what the system produces
Work directly with editorial partners on assigned projects as a primary point of contact for AI output quality, helping them understand what AI can and can't do and what is needed to improve results
Communicate the nuances of working with LLMs — including their stochastic nature, cost considerations, and the importance of evaluation — to engineering and editorial teammates in clear, practical terms
Contribute to documentation of prompting patterns, evaluation approaches, and quality standards that the team can build on over time
Stay current with developments in prompt engineering, LLM capabilities, and AI evaluation practices