Role Description Operationalizing Large Language Models requires specialized expertise beyond traditional MLOps practices. LLMs present unique operational challenges including significantly larger computational requirements, complex data pipelines, specialized infrastructure needs, and unique performance optimization requirements. This specialized role ensures GenAI solutions can scale effectively from proof-of-concept to enterprise-wide deployment in a utility environment. - Ensures GenAI solutions move successfully from prototype to production with proper operational support - Establishes specialized monitoring for model performance, inference latency, and data quality - Enables efficient scaling of LLM solutions across multiple business units - Creates high-performance deployment architectures that balance speed, cost, and reliability - Develops operational data pipelines to continuously improve model performance with new utility-specific data Qualifications - DevOps + ML: Expertise in Kubernetes, Docker, CI/CD tools, and MLflow or similar platforms - Cloud & Infrastructure: Understanding of GPU instance options, cloud services (AWS/Azure/GCP), and optimization techniques - Automation: Proficiency in Python, Bash, and infrastructure-as-code tools like Terraform or Ansible - LLM-Specific Frameworks: Experience with tools like TensorBoard, MLFLow, or equivalent for scaling LLMs - Performance Optimization: Knowledge of techniques to monitor and improve inference speed, throughput, and cost - Collaboration: Ability to work effectively across technical teams while adhering to enterprise architecture standards Requirements - Design and implement LLM-specific deployment architectures with Docker containers for both batch and real-time inference - Configure GPU infrastructure on-premises or in the cloud with appropriate CI/CD pipelines for model updates - Build comprehensive monitoring and observability systems with appropriate logging, metrics, and alerts - Implement load balancing and scaling solutions for LLM inference, including model sharding if necessary - Create automated workflows for model retraining, versioning, and deployment - Optimize infrastructure costs through intelligent resource allocation, spot instances, and efficient compute strategies - Collaborate with PG&E's Cyber team on implementing appropriate security controls for GenAI applications - Develop automated testing frameworks to ensure consistent output quality across model updates

Role Description The Senior Machine Learning Engineer will be responsible for: - Building recommendation and search across feed, discovery, search, and content continuation. - Owning retrieval/ranking: candidate generation, embeddings, two-tower models, features, and serving quality. - Designing, launching, and analyzing recommendation/search experiments. Qualifications - 5+ years industry experience building production ML systems with senior ownership. - Bachelor's degree from a recognized university in China is required. - Hands-on experience with recommendation, search, ranking, ads ranking, feed ranking, or content discovery systems. - Experience with consumer apps, entertainment, social, gaming, creator, or engagement-driven products. - Familiarity with two-tower models, embedding retrieval, candidate generation, ranking, and online/offline evaluation. Requirements - 5+ years industry experience building production ML systems with senior ownership. - Hands-on experience with recommendation, search, ranking, ads ranking, feed ranking, or content discovery systems. - Experience with consumer apps, entertainment, social, gaming, creator, or engagement-driven products. - Familiarity with two-tower models, embedding retrieval, candidate generation, ranking, and online/offline evaluation. Benefits - Salary range: 150k-300k+ equity. - This is a remote position. Anti-signals - Cannot show core Senior Machine Learning Engineer, Recommendation experience. - Not comfortable with the listed work mode. - Low ownership, coordination-only, or no shipped examples.

• Lead and architect IDP (Intelligent Document Processing) solutions: develop, train, and optimize robust pipelines for extracting unstructured data from complex documents, defining efficient fine-tuning, RAG, and computer vision strategies. • Architect Agentic AI development: design and productionize complex agentic workflows using LangGraph, n8n, and others, ensuring state management, resilience, low latency, and failure handling in the orchestration of the formalization and credit pipeline. • Model structured data for new AI products: build statistical pipelines and advanced analytical models on the data extracted by IDPs, transforming raw data into strategic business reports, market indices, and predictive intelligence. • Ensure Quality, Observability, and Cost (Evals & FinOps): structure automated evaluation frameworks (LLM Evals) to measure hallucination, accuracy, and recall of models in production, optimizing token consumption and infrastructure costs. • Autonomy and cross-functional collaboration: work directly with Product Managers, Engineering, and Technical Leadership to translate market pain points into viable architectures, participating actively from PoC design through final production delivery. • Mentorship and technical culture: support the technical development of more junior team members, disseminating best practices for coding, modeling, and AI usage.

• Support the development of IDP (Intelligent Document Processing) solutions: Contribute to creating, training, tuning models and fine-tuning for extracting unstructured data from complex documents. • Build the future of agentic AI: Assist in orchestrating workflows using frameworks such as LangGraph, n8n, and others, connecting document extraction to approval and formalization pipelines. • Transform data into products: Apply statistical techniques and data analysis to the structured datasets produced by our IDPs, helping identify patterns and insights that will become new AI products for our clients. • Collaborate with technical, product and commercial teams: Work side by side with Technical Leadership, Data Engineering, Product Managers and the Front-end team, evaluating, supporting and improving the transition of proofs of concept (PoCs) to production solutions. • Ensure quality and monitoring: Help evaluate model outputs (LLMs, RAGs and classical models), test prompts and document processes, ensuring solutions are accurate and scalable. • Learn and innovate: Stay up to date with the state of the art in Generative AI, bringing curiosity, ideas and proactivity to solve bottlenecks in our document pipeline.