AI Engineer, Generative AI Agents

We require people to be on-site, 4 days/week at our Denver or NYC office and are unable to offer relocation support.

About LG Ads:

LG Ads is a leading global ad solutions company, transforming the advertising landscape through innovative technologies. We are building the future of AdTech with our A.I. Data System (ADS), our proprietary AI agent designed to enhance every aspect of our business, from internal operations to external partner experiences.

Job Summary:

LG Ads is seeking a motivated AI Platform Engineer to join our AI team and help build the in-house platform powering generative AI agents deployed business wide. You'll contribute to the end to end stack; LLM serving and inference, RAG pipelines, evaluation harnesses and the APIs and infrastructure that put agents in production driving efficiency, unlocking new revenue streams and enabling self‑service data capabilities across the organization.

The role is hands‑on: you'll write production Python, integrate with services like AWS Bedrock and vector stores, instrument systems for observability and tune for latency, throughput and cost. We're looking for someone with a foundation in platform engineering and a working understanding of intent detection, query decomposition and context engineering, who is comfortable with agile delivery and can thrive in an ambiguous, fast‑paced environment.

• Build and maintain LLM-powered services and APIs (FastAPI, webhooks, Lang Graph), translating prototypes into production‑ready endpoints with proper error handling, retries and timeouts.
• Develop evaluation harnesses and offline/online eval pipelines to measure quality regressions, hallucination rates and task specific accuracy as models and prompts evolve.
• Instrument services with logging, tracing, and metrics (latency percentiles, token usage, error rates) so production behavior is observable and debuggable.
• Design and develop intelligent AI agents capable of intent recognition and decomposing complex queries into smaller, executable tasks that run in sequence or parallel.
• Implement and optimize context engineering techniques to ensure agents leverage relevant short and long‑term memory, as well as our aggregated knowledge base, for accurate and insightful responses.
• Integrate AI agents with internal systems such as ACR, Mosaic, and Salesforce, and third‑party services like Spring Serve and DSPs.
• Utilize and contribute to the development of standardized tooling protocols to streamline integration and maintenance of AI agents.
• Collaborate with cross‑functional teams, including product, and business units, to identify and build AI solutions that span the entire company.
• Develop and implement solutions for operational efficiencies, such as automating media planning and integrating agents into Mosaic (Home Grown DSP).
• Build client‑facing tools for voice and natural language data queries, supporting custom data and contributing to data monetization efforts for the ACR platform.
• Automate repetitive tasks in General & Administrative (G&A) functions, starting with Finance, and expanding to HR and IT Operations.
• Enable self‑service data access and analysis using AI agents, supporting diverse data sources.
• Participate in agile development sprints, actively contributing to planning, execution, and review.
• Manage ambiguity and adapt to evolving requirements in a rapidly developing AI landscape.

• Proven experience in designing and developing Generative AI agents.
• Strong expertise in intent detection and complex query decomposition.
• Demonstrated experience with context engineering for AI models.
• Proven understanding of knowledge graphs and/or vectorization as it relates to LLMs
• Experience building RAG systems including low level design and implementation of such systems.
• Solid understanding of large language models (LLMs) and their application in enterprise solutions, including foundational models like Llama 4 Maverick, Claude, and OpenAI.
• Experience with cloud platforms, particularly AWS, for scalable compute and storage resources.
• Familiarity with Databricks for processing structured and unstructured data, and its use as a runtime for…