Job Description
We are looking for a skilled Prompt Engineer to join our AI & Intelligent Automation team. In this role, you will design, test, and optimize prompts and AI workflows that power enterprise applications across customer service, field operations, regulatory compliance, and internal productivity. You will work primarily with Anthropic Claude (via AWS Bedrock), Microsoft Copilot, and supporting AWS infrastructure to build robust, production-grade AI solutions tailored to the utility industry.
This is a high-impact, cross-functional role — you will collaborate with business analysts, IT architects, data scientists, and frontline operational teams to translate complex utility use cases into reliable AI outputs.
Key Responsibilities
Prompt Design & Optimization
- Design, iterate, and maintain prompt libraries for AWS Agent Core, Claude (Anthropic) and Microsoft Copilot across enterprise use cases including outage communications, meter data analysis, regulatory filings, and customer inquiries.
- Apply advanced prompting techniques — chain-of-thought, few-shot learning, retrieval-augmented generation (RAG), and tool use — to improve output quality and consistency.
- Build and maintain system prompts, instruction sets, and guardrails aligned with client brand, safety, and compliance standards.
- Conduct A/B testing and prompt regression analysis to benchmark output quality over time.
AWS & Infrastructure Integration
- Deploy and manage AI workloads on AWS Bedrock, leveraging Claude API integrations, model fine-tuning pipelines, and serverless architectures (Lambda, API Gateway, S3).
- Collaborate with cloud architects to build scalable, cost-efficient prompt orchestration pipelines.
- Implement logging, monitoring, and alerting for AI model performance using AWS CloudWatch and related tooling.
- Ensure data privacy and security compliance (CPUC, FERC, SOC 2) in all AI workflow designs.
Microsoft Copilot Deployment
- Configure and extend Microsoft Copilot across M365 environments (Teams, Outlook, SharePoint, Power Platform) for field crews, customer service reps, and executive stakeholders.
- Develop Copilot plugins and connectors integrating with enterprise internal systems (SAP, Salesforce, GIS platforms).
- Train internal users and champions on effective Copilot usage; develop prompt guides and self-service resources.
Collaboration & Governance
- Partner with legal, compliance, and cybersecurity teams to document AI use cases, risk assessments, and model governance artifacts.
- Participate in the AI Ethics and Responsible Use working group, ensuring equitable and transparent AI deployment.
- Document prompt engineering best practices, version-control prompt libraries in Git, and contribute to internal knowledge base.
- Stay current on model releases (Claude 3/4, GPT-4o, Gemini, etc.) and evaluate applicability to utility operations.
Qualifications
Required
- 3+ years of experience in prompt engineering, AI/ML engineering, data science, or a closely related field.
- Hands-on experience with large language models — particularly Anthropic Claude — in production environments.
- Proficiency with AWS services: Bedrock, Lambda, S3, IAM, CloudWatch, and API Gateway.
- Experience configuring and deploying Microsoft Copilot in enterprise M365 environments.
- Strong written and verbal communication skills — able to translate technical concepts for non-technical stakeholders.
- Familiarity with Python and/or JavaScript for scripting, API calls, and automation.
- Understanding of data privacy regulations relevant to utilities (CPUC Rule 25, etc.).
- Experience with prompt engineering tools, frameworks (LanChain, LlamdaIndex and evaluation methods.
Preferred
- Experience in the energy, utility, or regulated infrastructure sector.
- Exposure to RAG architectures, vector databases (Pinecone, OpenSearch), or LangChain/LlamaIndex.
- AWS Certified Machine Learning Specialty or AWS Solutions Architect certification.
- Microsoft Certified: AI Engineer Associate or Copilot-related certification.
- Familiarity with SAP S/4HANA, Salesforce, or GIS platforms common in utility operations.
- Knowledge of NERC CIP, CPUC, or FERC regulatory frameworks.