GenAI Workflow & Service Development

Design, implement, and maintain Python-based services and workflows that integrate LLMs and GenAI capabilities with internal platforms and applications.
Build and iterate on agentic and multi-step workflows using approved orchestration frameworks and platform patterns (e.g., LangGraph, AgentCore, LangChain).
Consume existing retrieval/RAG and search abstractions to improve response quality, grounding, and reliability, tuning parameters (top-k, scoring, filters) rather than re-implementing core retrieval infrastructure.
Develop robust tooling and APIs for agents, including clear input/output schemas, error contracts, versioning, and observability hooks.
Implement structured output handling (e.g., JSON schemas, tool calls) to ensure predictable behavior and easier downstream integration.

Platform Integration & Governance

Operate within established platform, security, and governance guardrails (RBAC, data access boundaries, PII handling, logging, audit) instead of building custom, one-off mechanisms.
Leverage existing platform SDKs, templates, and patterns for configuration, deployment, and monitoring of GenAI workloads.
Work in a cloud-native AWS and Azure environments (e.g., cloud native applications, environment variables, secrets management, logging/metrics/tracing), collaborating with platform teams as needed rather than owning core infra design.

Partner with product managers, internal business stakeholders, and UX to translate problem statements and evaluation criteria into concrete, production-ready workflows.
Collaborate with data and application teams to integrate GenAI capabilities into existing systems (e.g., internal tools, portals, automation flows) with minimal disruption.
Participate in design reviews, code reviews, and architecture discussions, ensuring solutions are maintainable, observable, and aligned with platform standards.
Contribute to internal enablement (playbooks, examples, patterns) to help eTech Engineering and other teams become effective consumers of GenAI capabilities.

Own the operational health of GenAI workflows you build: monitoring, alerting, troubleshooting, and iterative improvement.
Incorporate evaluation and guardrail checks (e.g., automated tests, evaluation harnesses, red/blue team feedback) into workflows to improve quality and safety over time.
Act as a “high adopter of AI to build AI”, continuously using AI tools (e.g., Glean, Devin, Windsurf, Claude) to accelerate design, development, testing, and documentation.