12 months ago
Responsibilities
- Architect and build agentic AI systems that plan, reason, and act across workflows.
- Design tool-using agents integrated with Powerline systems.
- Develop multi-step decision systems and stateful, long-running agents.
- Define and implement evaluation frameworks for agents.
- Improve the reliability, observability, and safety of LLM-based systems in production.
- Collaborate with energy-market specialists to translate real-world workflows into agent behavior and logic.
- Help lead technical direction for agent architecture and system design.
- Build AI agents that answer complex questions, explain algorithm results, provide decision support, and run continuous autonomous tasks.
Requirements
- Experience building production AI or software systems, including LLM-powered or agentic applications.
- Strong Python skills and experience integrating with APIs and external systems.
- Experience designing multi-step or stateful systems such as pipelines, workflows, or agents.
- Experience building memory and Retrieval Augmented Generation (RAG) frameworks.
- Experience evaluating agents using direct user feedback.
- Experience with agent orchestration frameworks such as Pydantic AI, LangChain, or similar.
- Ability to work across engineering, product, and applied AI.
- Strong ownership and comfort building 0-to-1 systems in fast-paced environments.
- Excellent communication and collaboration skills across technical and domain teams.
- Experience building systems with partial or full autonomy in real-world environments is a strong plus.
- Experience with real-time or streaming data pipelines is a strong plus.
- Experience in electricity markets, trading, or energy systems is a strong plus.
- Experience with time-series modeling or forecasting is a strong plus.
- Experience with optimization systems is a strong plus.
Tech Stack
Categories
About Powerline
Powerline builds AI-driven software for monitoring and operating energy assets, with a focus on optimizing renewable generation and battery storage in power markets. Its platform ingests real-time data and applies machine learning and optimization to improve dispatch, revenue, and emissions outcomes for asset owners, grid participants, and energy traders. The privately held company is based in the San Francisco Bay Area.