About the job
Joist AI develops AI-driven software for the architecture, engineering, and construction (AEC) sector. The team aims to simplify collaboration and help organizations improve marketing and revenue operations through intelligent technology.
Role overview
This remote Agentic Systems Engineer position is open to candidates based in India. The focus is on building smart applications that streamline proposal writing for AEC professionals. The work involves multi-agent orchestration, MCP servers, skills integration, memory management, evaluation systems, retrieval tools, and production infrastructure. Candidates should have 2–4 years of experience and a strong interest in creating tools that reason, remember, and collaborate with users.
What you will do
- Design and develop modular agents that integrate into a larger system architecture.
- Implement memory layers, including short-term, long-term, summarization, and retrieval-backed components.
- Integrate tools, MCP servers, and skill modules into the system.
- Test, evaluate, and monitor features to ensure quality before release.
- Review production traces to understand system behavior and address issues as they come up.
Requirements
- 2–4 years of experience building production software.
- Strong Python skills, including the ability to spot high-quality code and identify issues in LLM-generated code. Comfortable with clean object-oriented programming, idiomatic syntax, modular design, and thorough testing.
- Solid understanding of agentic and LLM concepts such as RAG, prompting, tool use, structured outputs, streaming, context management, and generative AI basics.
- Hands-on experience with modern agent toolkits, gained through projects, prototypes, or hackathons.
- Ability to read and understand unfamiliar codebases quickly.
- Detail-oriented approach to troubleshooting and problem-solving.
- Decisions informed by data from production traces, evaluation metrics, and logs.
