Principal Generative AI Architect- 15+ yrs- Los Angeles, United States- Onsite

Key ResponsibilitiesGenAI Strategy & Platform Architecture:Define client’s multi-year Generative AI blueprint, model selection strategies (open-source vs. proprietary), and framework governance.Design enterprise-grade RAG architectures utilizing high-throughput vector databases (Pinecone, Milvus, pgvector) connected to video telemetry, content metadata libraries, and customer data platforms.Architect multi-agent orchestration systems (LangGraph, AutoGen, Semantic Kernel) for complex, multi-step workflows.Infrastructure & Cost Optimization:Design cost-efficient, low-latency inference pipelines on cloud infrastructure (AWS Bedrock, Azure OpenAI, Google Cloud Platform Vertex AI).Implement GPU utilization strategies, model quantization, caching layers, and token usage optimization to manage enterprise compute costs.Security, Ethics & Governance:Establish enterprise guardrails, hallucination detection mechanisms, toxicity filters, and data loss prevention (DLP) for LLM interactions.Ensure compliance with digital rights management (DRM), user privacy (CCPA), and copyright protections when ingesting video scripts, captions, and user logs.Cross-Functional Leadership:Partner with Product Managers, Data Engineering teams, and Content Operations to turn business ideas into viable AI initiatives.Guide GenAI Engineers through technical mentorship, code reviews, and architecture decision records (ADRs).Qualifications & RequirementsEducation: Master’s or Ph.D. in Computer Science, Artificial Intelligence, Machine Learning, or related field (or equivalent practical experience).Experience: 10+ years in software engineering/data architecture, with 3+ years specifically leading Generative AI architecture, LLM deployment, and RAG systems at scale.Technical Stack: Deep expertise in Python, PyTorch/TensorFlow, LangChain/LlamaIndex, Vector DBs, Model Serving (Triton, vLLM), and Cloud AI Platforms (AWS/Google Cloud Platform/Azure).Domain Expertise: Familiarity with recommendation engines, conversational search, multi-modal systems (text-to-video/audio-to-text), or media workflows is a major plus.