Prologis is looking to leverage AI to improve the efficiency and effectiveness of its operations, specifically by developing an AI agent for the CTO office to assist with tasks like meeting preparation, decision memos, and knowledge retrieval. The goal is to automate processes, quantify business value, and integrate AI solutions securely with enterprise systems.
Requirements
- technical fluency required (you can read/write basic code and talk data).
- Hands-on with Python and SQL; call REST APIs, parse JSON
- Working knowledge of LLMs and agent patterns: RAG, tool/function calling, planning, memory, and evaluation
- Experience with OpenAI APIs, prompt design
- Analytics basics: build a small dashboard (Tableau/Sigma) and compute straightforward product metrics.
Responsibilities
- Ship an MVP Chief-of-Staff AI agent for the CTO office within 10–12 weeks (meeting prep/briefs, decision memos, action-item capture, stakeholder Q&A, knowledge retrieval across Outlook/Teams/SharePoint/Confluence/Jira).
- Design agent workflows (tool use/functions, RAG, multi-agent orchestration, memory, guardrails) and instrument success with clear KPIs (task success rate, precision/recall for retrieval, latency, and cost/run).
- Map processes (BPMN or equivalent) to find automation candidates; quantify business value (hours saved, cycle-time reduction, error-rate deltas) and prioritize a backlog with acceptance criteria.
- Integrate securely with enterprise systems and partner with Security for reviews
- Build end-to-end prototypes: data prep (Python/SQL), agent logic (Dataiku or similar), and lightweight UIs for demos.
- Run fast experiments: with documented results.
- Communicate crisply: weekly 1-page status, exec-ready demo scripts, and an operational run book so IT can support what you build after the internship.
Other
- Currently enrolled in an MBA or MS program
- Business skills: requirements gathering, prioritization, building a clear backlog from ambiguity, and delivering to a deadline.
- Communication: concise writing (1-pagers, READMEs) and confident exec-level presentations.
- Execution style: bias to action, iterate quickly, seek feedback, and manage your own project plan.
- Individuals take the initiative here.
