Software Engineer - Agentic AI Platform
Software ENG & DEV | San Francisco, CA | Full Time, Contract, and Temporary | From $75.00 to $100.00 per hour
Job Description
Software Engineer - Agentic AI Platform
- Hourly pay: $75-$100/hr (Depends on years of experience)
- Worksite: Leading audio, video, and voice technologies company (Remote - Open for candidates located in the United States)
- W2 Employment, Group Medical, Dental, Vision, Life, Retirement Savings Program, PSL
- 40 hours/week, 6 Month Assignment
A leading video, audio, and voice technologies company is seeking a Software Engineer - Agentic AI Platform to build and extend a centralized Agentic AI framework that enables secure, governed, and cost-aware AI agents across enterprise GitLab workflows. This role will focus on developing AWS-based agent orchestration services, integrating AI capabilities into CI/CD pipelines, implementing security and observability controls, and delivering scalable platform services that enable engineering teams to adopt AI without rearchitecting their existing
Software Engineer - Agentic AI Platform Responsibilities:
- Design, develop, and maintain AI agent orchestration services using AWS Lambda, SQS, EventBridge, API Gateway, and Amazon Bedrock Agents; build routing logic, event-driven workflows, and scalable agent execution frameworks.
- Integrate AI agents into GitLab CI/CD pipelines by developing reusable pipeline patterns, automating agent execution stages, consuming repository and testing context, generating artifacts, and enforcing pass/fail quality gates.
- Build and maintain Bedrock Action Groups, Knowledge Bases, and retrieval systems leveraging AWS Lambda, OpenSearch, S3, and enterprise data sources to enable contextual and scalable AI decision-making.
- Implement security, governance, and operational controls, including Bedrock Guardrails, IAM policies, secrets management, input/output validation, audit logging, token usage controls, model routing strategies, semantic caching, and cost optimization mechanisms.
- Develop observability, monitoring, and reporting solutions using CloudWatch and related AWS services; create dashboards, tracing, logging, compliance reporting, and operational insights while partnering with engineering teams to deliver reliable AI platform capabilities.
Software Engineer - Agentic AI Platform Qualifications:
- 3-7 years of professional software engineering experience developing cloud-native applications and distributed systems.
- A bachelor's degree in Computer Science, Software Engineering, Information Technology, or a related technical discipline is preferred.
- Strong Python development experience, including AWS SDK (Boto3) and production-grade AWS Lambda development.
- Hands-on experience with AWS services, including Lambda, API Gateway, SQS, EventBridge, IAM, Secrets Manager, and CloudWatch.
- Experience working with Amazon Bedrock, including Agents, Action Groups, Knowledge Bases, and Guardrails.
- Experience integrating with LLM platforms such as Amazon Bedrock, OpenAI, Anthropic, or similar AI services.
- Strong understanding of GitLab CI/CD pipelines, automation workflows, and software delivery practices.
- Experience designing secure cloud solutions with knowledge of IAM, secrets management, least-privilege access controls, input validation, and AI security considerations such as prompt injection and data protection.
- Experience implementing observability solutions, including structured logging, monitoring, tracing, and production troubleshooting.
- Experience developing event-driven architectures, APIs, and microservices is preferred.
- Experience with OpenSearch, vector search, embeddings, retrieval-augmented generation (RAG), or semantic search technologies is preferred.
- Experience with FinOps, TokenOps, model routing, semantic caching, cost attribution, or AI operational governance is preferred.
- Experience building internal developer platforms, engineering enablement tools, or shared infrastructure services is preferred.
- Familiarity with agent frameworks and multi-agent architectures is preferred.
#PP
