Role Overview
This is a project-focused internship for an AI engineer embedded on the DevOps team. You will report to the DevOps lead and partner with internal technology teams. The work centers on internal, production-adjacent tooling—not training or shipping customer-facing ML models.
Core Internship Projects
- Generative assistant for alert response: Learn the observability stack (Prometheus, Grafana, Loki, Tempo, Alertmanager, OpenTelemetry) and prototype a generative agent that proposes structured mitigation suggestions for alerts with traceability back to queries, dashboards, or signals.
- Retrieval on internal data (RAG): Build and iterate on RAG over permissioned internal data sources (runbooks, tickets, docs, system design, postmortems) so suggestions and Q&A are grounded and citeable.
- Path toward agentic remediation: Outline how the system could execute approved remediations behind explicit guardrails and human approval. Implement only what is allowed and under review.
- Broader internal Q&A: Explore how additional internal, permissioned firm data can support natural language questions for engineers.
Responsibilities
- Design and prototype agent workflows with tool use, policy boundaries, and human-in-the-loop where appropriate.
- Collaborate with platform and service teams to make more observability and operational context available in a safe, governed way for agents.
- Document experiments, limitations, evaluation approach, and safety assumptions; ship changes via Git (branches, merge requests, meaningful commits).
Requirements
- Pursuing a BS or MS in Computer Science, Computer Engineering, Information Systems, or a related field.
- Expected to graduate by Summer 2028.
- Hands-on experience using AI tools (e.g. LLM APIs, assistants, or coding agents) in real projects; preferably experience building an agent.
- Experience with RAG (retrieval design, chunking, evaluation, grounding, or production-minded prototyping).
- Python for prototyping and integration; comfortable consuming and creating APIs and working with JSON; able to understand YAML for configs.
- Git workflow: branches, merge requests, meaningful commit messages.
- Strong judgment on data handling: no secrets in prompts/logs, minimize sensitive data, follow internal policies.