The RoleWe are looking for an AI Engineer to build the agentic systems at the core of Jeevy.
You will design and ship AI agents that interact with our ERP, reason over engineering and manufacturing data, call tools, maintain project context, and execute multi-step workflows across quoting, project management, scheduling, procurement, quality, and production.
This is an applied engineering role. You will work directly with fabrication shops and Jeevy's internal operations team, observe where work breaks down, and turn those workflows into software.
Examples of systems you might build:
- An estimator that reads engineering drawings, BOMs, specifications, and historical jobs to produce a first-pass quote.
- A project agent that converts a purchase order into deliverables, work orders, schedules, procurement requirements, and deadlines.
- A production agent that monitors shop-floor progress and identifies when a job is likely to miss its schedule.
- Agents that reconcile information across drawings, emails, ERP records, quality documents, and customer requirements.
- Verification systems that prevent agents from taking actions unless their conclusions can be traced back to source documents or system state.
- Long-term memory and knowledge systems that allow Jeevy to understand how a shop operates over thousands of projects.
You will own these systems from architecture through deployment.
What You'll Do- Design and build production-grade AI agents for complex industrial workflows.
- Build tool-calling systems that allow models to safely interact with ERP data and operational systems.
- Develop agent architectures involving routing, planning, specialized skills, memory, and verification.
- Build structured extraction pipelines for engineering drawings, specifications, BOMs, PDFs, emails, images, and other unstructured data.
- Design evaluation systems for measuring agent correctness, reliability, and task completion.
- Develop retrieval and knowledge architectures for project- and shop-level memory.
- Build guardrails and human-in-the-loop systems for high-consequence actions.
- Improve model latency, reliability, cost, and context management.
- Work directly with users in fabrication shops to understand real workflows and rapidly deploy improvements.
- Contribute across the stack when necessary to ship complete products rather than isolated AI prototypes.
What We're Looking ForYou should be an exceptional software engineer who has gone deep on modern AI systems.
Strong candidates will have experience with several of the following:
- Python and production backend systems.
- LLM APIs and modern foundation models.
- Agentic systems, tool use, planning, routing, or multi-agent architectures.
- Structured outputs and constrained generation.
- RAG, embeddings, semantic search, and knowledge retrieval.
- Evaluation frameworks for probabilistic AI systems.
- Document understanding, multimodal models, OCR, or vision-language models.
- Graph-based agent orchestration such as LangGraph or similar systems.
- SQL and production databases.
- Building reliable systems around inherently unreliable model outputs.
You do
not need a manufacturing background. You do need to be excited about learning how physical things get built.
You Might Be a Great Fit If- You care about deploying rapidly and iterating quickly to make exceptional products
- You naturally think about state, tools, permissions, failure modes, verification, and observability when designing agents.
- You are comfortable operating in ambiguous environments where the correct product architecture has not yet been discovered.
- You want to sit with real users, watch them work, and ship software immediately afterward.
- You enjoy moving between AI research papers, backend architecture, product design, and messy real-world operational problems.
- You want your software to control or coordinate something tangible in the physical world.
- You prefer high ownership and fast iteration over narrowly scoped engineering work.
Bonus- Experience with manufacturing, aerospace, construction, industrial automation, logistics, or ERP systems.
- Computer vision or multimodal ML experience.
- Experience working with engineering drawings or CAD data.
- Knowledge graphs or graph databases.
- Optimization, scheduling, or operations research.
- Fine-tuning, synthetic-data generation, or model training.
- Experience deploying AI systems in environments with strict security, compliance, or data-governance requirements.
Location: San Francisco, Fort Worth, Remote for Exceptional Canddiates
Type: Full-time
