Descripción
Full-Stack Engineer (Python + React + AI Agents)
Ingram Micro is seeking an embedded full-stack engineer who will deliver features across: agent prompts and tools, FastAPI services, Postgres/SQLite persistence, and the React audit-review UI. Each engineer is expected to be comfortable moving between the agent layer, the API layer, and the front-end inside a single feature.
Key responsibilities
- Build and maintain features in the Python/FastAPI backend (server, routes, services) including audit orchestration,
audit cache, scheduler, reporting, and integrations with internal Ingram applications, third-party SaaS tools, and
infrastructure systems.
- Implement and refine agents and prompts: output_schema-bound verdict agents, MCP tool wiring, prompt templates
and SOP-injection blocks.
- Build and harden React/TypeScript components in the audit-review UI: report views, verdict drill-downs, override
workflows, and dashboards consuming the Insight Agent output.
- Write Pydantic models, schema-repair logic, and unit/integration tests.
- Integrate with change-management, HR, and identity/directory systems through MCP-bound tool servers and REST
clients.
- Contribute to the Postgres / SQLite data model and to migration / backfill scripts.
- Participate in code reviews, on-call triage, and accuracy/regression reviews.
- Ship features with the platform's accuracy controls already in place: temperature 0.0 on verdict agents, mandatory
evidence citation, deterministic prompt construction, and reproducibility checks.
What You Bring:
- 4+ years of full-stack engineering, with strong production experience in BOTH Python and React/TypeScript.
- Python: FastAPI (or Flask/Starlette), Pydantic v2, asyncio, SQLAlchemy, Postgres, REST API design, pytest.
- React/TypeScript: functional components, hooks, state management, typed API clients, working with paginated
tables and detail views.
- Hands-on experience building or integrating AI agents - at least one production or substantial prototype using Google
ADK, LangGraph, LangChain, OpenAI Assistants, or comparable, including tool-use and structured outputs.
- Comfortable with prompt engineering for frontier LLMs (Gemini, Claude, GPT-4-class) and with debugging
structured-output failures.
- Git, code review, and trunk-based / feature-branch workflows.
- Docker and basic familiarity with containerized cloud deployments.
Nice-to-have
- MCP (Model Context Protocol) server or client experience.
- Experience integrating with change-management, HR, and identity/directory systems via REST APIs or equivalent.
- SQLite + Postgres dual-store applications.
- Bitbucket Pipelines or equivalent CI/CD experience.
- Exposure to GRC / IAM / SOX / ITGC tooling.
- Prior work on LLM evaluation harnesses and regression testing.
