About the Role
Job Title: Software Engineer – Agentic AI PlatformLocation: Remote (Client location: 1275 Market Street, San Francisco, CA)Duration: 6 months
About the RoleWe're building a Centralized Agentic AI Framework, a shared platform that brings safe, governed, and cost-aware AI agents into every team's GitLab workflow.You will be a hands-on engineer building and extending this platform by developing Lambda functions that orchestrate agents, integrating Amazon Bedrock with GitLab CI/CD, strengthening security and observability layers, and delivering new agent capabilities that teams can adopt without re-architecting their workflows.
Responsibilities• Build agent orchestration using Router Lambdas, SQS-based queuing, and Amazon Bedrock Agent invocation.• Integrate with GitLab CI/CD by designing .gitlab-ci.yml patterns where agent invocations run as pipeline stages, consume branch, diff, and test context, generate artifacts, and gate downstream stages based on agent pass/fail signals.• Develop shared Action Groups using AWS Lambda-backed tools.• Design Knowledge Bases using Amazon Bedrock Knowledge Bases, OpenSearch, and S3 to enable agents to reason over shared organizational context.• Implement Bedrock Guardrails, including input/output filtering, sensitive-data scrubbing, content filtering, and per-agent permission boundaries.• Implement TokenOps controls, including model tiering, model routing through an LLM gateway, semantic caching, and context-window management.• Build observability solutions using CloudWatch dashboards, audit trails, and trace logging for agent invocations, including inputs, outputs, decisions, token usage, and cost tracking.
Required Qualifications (3–7 Years Experience)• Strong Python skills and experience with AWS SDK (Boto3) for building production Lambda functions and event-driven services.• Hands-on experience with AWS services including Lambda, API Gateway, SQS, EventBridge, IAM, Secrets Manager, and CloudWatch.• Experience with Amazon Bedrock, including Agents, Action Groups, Knowledge Bases, and Guardrails.• Experience integrating with LLM APIs such as Amazon Bedrock, OpenAI, Anthropic, or similar platforms.• Familiarity with CI/CD platforms, preferably GitLab CI.• Strong security mindset, including least-privilege IAM, secrets management, input validation, and awareness of prompt injection and data leakage risks.• Experience with observability practices including structured logging, metrics, tracing, and production debugging.Nice to Have• Experience with vector search and RAG solutions, including OpenSearch, embeddings, and retrieval evaluation.• FinOps or TokenOps experience, including cost attribution, model routing, semantic caching, and batch inference.• Experience building developer platforms or internal tools used by engineering teams.• Familiarity with agent frameworks such as Bedrock AgentCore and their trade-offs.How You Work• Write small, focused services with clear contracts.• Design for extensibility so new triggers, agents, and tools can be added without impacting unrelated components.• Focus on developer experience as well as technical implementation.• Comfortable working with ambiguity and debugging non-deterministic agentic systems through empirical analysis.