ProDex Labs Inc Senior AI Engineer New York, NY · Full time Company website

Apply for Senior AI Engineer

ProDex Labs builds Dexter, an AI industrial engineer for factory operations. Dexter connects enterprise systems, spreadsheets, and operational knowledge into a live factory model that manufacturers use to simulate production, optimize schedules, identify bottlenecks, and reduce quality costs. Our customers include L3Harris, HighCom Armor Solutions, TIME Manufacturing, and Alimentos ProSalud. We are a small team in Manhattan building software for demanding industrial environments, with deployment inside customers’ security boundaries.

About the Role

We are hiring a Senior AI Engineer to lead the architecture and development of Dexter’s agent systems. You will build AI that reasons over complex factory data, coordinates analytical tools, and produces answers that manufacturing teams can inspect and use to make consequential operational decisions.

You will own the agent layer across simulation, scheduling, bill of materials, and cost modeling. This includes tool orchestration, retrieval, memory, and the evaluation systems that measure correctness and catch regressions. The work requires strong software engineering, a rigorous approach to validation, and sound judgment about where models need deterministic computation or human input.

As a senior individual contributor, you will set technical direction, lead architectural decisions, and raise the engineering standard through design reviews and mentorship. You will work directly with customers and the product team to translate unfamiliar manufacturing workflows into reliable capabilities, from initial design through production operation.

In this role, you will

• Architect Dexter’s agent systems, including planning, tool selection, memory, error recovery, and coordination across analytical engines.
• Build retrieval and data interpretation pipelines for structured records, spreadsheets, scanned documents, and operational knowledge, with clear source attribution and permission boundaries.
• Own evaluation infrastructure that measures task completion, numerical correctness, grounding, and failure modes using representative customer workflows and regression tests.
• Work directly with production planners and industrial engineers to validate assumptions, investigate failures, and turn customer feedback into product improvements.
• Make architecture decisions across model selection, inference, retrieval, and fine-tuning, balancing accuracy, latency, cost, and deployment constraints, including local inference in disconnected environments.
• Lead design and code reviews, mentor engineers, and establish release practices informed by production traces, automated testing, and measurable quality improvements.

What You Bring

• 7+ years of software engineering experience, including 2+ years building and operating production systems using large language models.
• Strong Python proficiency and experience designing maintainable services, APIs, and data-intensive applications.
• Experience building evaluation frameworks and using quantitative evidence to guide model, prompt, retrieval, or agent architecture changes.
• Deep understanding of agent tool use, retrieval, context management, and the failure modes of probabilistic systems in production.
• A track record of leading technically complex projects, making architectural tradeoffs, and improving the quality of work across a team.
• Ability to learn unfamiliar domains, translate ambiguous requirements into technical plans, and communicate decisions clearly to engineers and customers.

What Will Set You Apart

• Experience with operations research, discrete event simulation, mathematical optimization, or mixed-integer linear programming.
• Experience deploying models or agent systems in private, on-premises, or air-gapped environments.
• Experience building decision-support software where correctness, auditability, and operational consequences are central to product quality.

We encourage candidates with strong relevant experience to apply, including those whose backgrounds do not align with every qualification. Some customer assignments require U.S. person status under ITAR.

Help build the software that advances American manufacturing.