Program Manager

What you’ll do

Define and implement agent-customer communication standards.
Help define quality for user-facing support bots.
Conversational Designer for Uber Customer Support.
Work with Product and Engineering to update the Response Generation prompt.
Evaluate the performance of our agent interactions at scale.
Optimize support resolution steps for interaction efficiency.
Work with engineering teams to optimize how knowledge is used within a customer-support conversation.
Design prompt improvements for chatbots powered by LLMs.
Develop, own, and implement short & long-term strategies for agent responses in support interactions.
Work with our Metadata team to improve intent identification.
Continuously enhance all agent-response products to ensure improvement to key metrics & user efficacy.
Coordinate strategic projects between GDX, CO, and Global Content.
May telecommute.

Basic Qualifications

Employer will accept a Master's degree in Conversational Design, Business, Project Management, Information and Knowledge Strategy, or related field and 3 years of experience in the job offered or in a related occupation.

Position requires:
Prompt engineering and optimization for LLMs to optimize responses, design structured task flows, and improve model accuracy and efficiency in dynamic contexts;
Applying large language models (LLMs) to extract insights, automate processes, and generate high-quality natural language outputs for business or technical use cases;
Stakeholder management skills, including in delivering change management projects to launch new enterprise-wide tools or systems;
Building retrieval-augmented generation (RAG) chatbots with a vector database, similarity search-based reranking and retrieval, and response generation for combining proprietary data sources with generative AI models that ensures factual and contextually grounded outputs;
Model fine-tuning for customizing pre-trained LLMs on domain-specific datasets, improving performance on targeted tasks such as Q&A through at least one of the following: HuggingFace’s API,

OpenAI’s API;
Implementing Python code for data analysis, automation, API integration, and machine learning pipeline development, including libraries like Pandas, NLTK, Scikit-learn, or PyTorch, and utilizing cloud environments like Google Cloud Platform (GCP) for deploying and managing machine learning models, orchestrating data workflows, and scaling compute resources efficiently;
Qualitative analysis of text data, including thematic analysis and sentiment analysis to interpret customer feedback, identify emerging issues, and inform product, marketing, or support strategies;
Designing a knowledge management strategy, roadmap, and implementation framework to ensure that stakeholders can share and access the needed information and knowledge at the right time;
Designing a continuous improvement process through retrospectives, lessons learned capturing, and cycles of feedback integration;
Information Architecture for maintaining metadata, tagging, findability, and knowledge retrieval systems;
Information organization for organizing, structuring, and maintaining high-quality documentation and knowledge assets;
Understanding of knowledge base content curation and best practices, including background in content creation and curation, copy-editing and style guidelines that inform conversational design.