Y

Manager, AI Engineering - Taco Bell Voice AI

Yum! Brands
Full-time
On-site
Irvine
We are hiring a

Manager, AI Engineering

to lead the development of intelligent voice agent behaviors for the

Taco Bell Voice AI

program. This role sits at the intersection of LLM prompt engineering, real-time orchestration, and applied AI system design. You’ll help craft agent logic, deploy optimized LLM and ASR-based components, and ensure the overall responsiveness and quality of the drive-thru AI experience. This role is ideal for a senior AI practitioner who thrives in high-velocity environments and is excited by the technical and creative challenges of voice-based AI systems. Key Responsibilities: Agent Development & AI Orchestration:

Design, author, and maintain prompt chains, decision flows, and persona logic. Integrate LLMs with speech, menu data, POS systems, and structured knowledge sources. Lead the implementation of fallback logic, retry mechanisms, and safety filters. Model Fine-Tuning & Optimization:

Fine-tune LLMs, ASR, and embedding models to optimize agent behavior and performance. Leverage techniques like LoRA, quantization, distillation, and pruning. Collaborate with MLEs to measure and reduce latency, response failure, and hallucination rates. Team Leadership & Collaboration:

Mentor AI Engineers and Data Scientists on prompt strategy and model tuning. Define evaluation workflows and scalable agent experimentation pipelines. Work cross-functionally with PMs, MLEs, QA, and voice platform teams to align priorities. Required Qualifications: + years experience in AI Engineering, ML Systems, or LLM-driven development. Deep experience with LLMs ( OpenAI, Llama, Nemotron, Mistral) and prompt-based application design. Strong coding skills in Python and experience with LangChain, vector stores, and API orchestration. Experience fine-tuning or adapting ASR, embedding, or LLM models in production. Preferred Qualifications: Familiarity with real-time audio systems and multi-turn conversational agents. Experience deploying LLM agents in customer-facing use cases. Understanding of speech interfaces, retrieval-augmented generation (RAG), and evaluation metrics. Location & Travel: This role is based in Irvine, CA at the Taco Bell offices. Hybrid work model: in-office – days per week required. Occasional travel to test markets or partner locations may be required. Salary Range: $,-$, annually + bonus eligibility.

This is the expected salary range for this position. Ultimately, in determining pay, we'll consider the successful candidate’s location, experience, and other job-related factors.