You'll have the opportunity to publish at top-tier venues like NeurIPS, ICML, and ACLand see that research deployed to a user base of over 80 million customers.

Job Responsibilities

• Lead research and deployment of agentic AI systems with multi-step workflows, tool calling, and multi-agent orchestration.

• Fine-tune and optimize LLMs using parameter-efficient fine-tuning (PEFT), distillation, and quantization to meet production constraints such as latency, memory, and cost.

• Apply reinforcement learning and preference optimization to improve personalization and dialogue policies.

• Scale LLM systems through caching, batching, prompt governance, and evaluation frameworks.

• Implement privacy, safety, and security controls including PCI compliance, jailbreak resistance, and auditability.

• Design rigorous experiments with strong baselines and meaningful metrics.

• Define and track success metrics for agent performance, including task completion rate, accuracy, latency, and customer satisfaction.

Required Qualifications, Capabilities, and Skills

• Ph.D. with 1+ years or M.S. with 3+ years building and deploying AI systems in production

• Applied GenAI experience with LLMs including fine-tuning, prompt engineering, and RAG.

• Experience scaling LLM systems with caching, batching, governance, and evaluation.

• Strong foundation in ML, deep learning, statistical modeling, and experimental design.

• Experience in Information Retrieval (indexing, ranking, retrieval) and/or recommendation systems.

• Proficiency in Python and ML frameworks (PyTorch/TensorFlow, Hugging Face, scikit-learn)

• Demonstrated ability to set a technical research agenda and drive it from concept through production deployment.

• Experience presenting research findings and technical strategy to senior leadership and non-technical stakeholders.

Preferred Qualifications, Capabilities, and Skills

• 5+ years developing conversational AI systems, virtual assistants or LLM-based systems in production.

• Experience with multi-agent orchestration, supervisor agents, and specialized toolkits.

• Expertise in agent governance, red-teaming, adversarial testing, and safety evaluation.

• Experience with reinforcement learning, bandit algorithms, and preference-based optimization (DPO, IPO), with practical exposure to data collection, labeling, and evaluation pipelines.

• MLOps/LLMOps experience with CI/CD, monitoring, versioning, A/B testing, and rollbacks.

• Track record of data-driven product development and experimentation.

• Publications in top-tier AI/ML venues and/or open-source contributions

You'll ship production LLM systems to millions of customers, publish in top-tier venues, and solve hard problems in agent accuracy, reliability, and safety.