Founding Engineer - AI/ML at Manufex
About Manufex
Manufex is a US-based on-demand manufacturing marketplace connecting hardware and industrial companies with vetted suppliers across CNC machining, sheet metal fabrication, 3D printing, and more. We bring pricing transparency, quality assurance, and streamlined procurement to a market historically run on emails and phone calls. The core of what makes Manufex different isn't the marketplace it's the intelligence layer sitting on top of it. Every order on Manufex runs through a unified thread powered by a suite of AI agents and a PM that manages order health end-to-end. We're building the operating system for modern manufacturing and we're looking for the engineer who will build the AI that powers it.
About the Role
This is a founding engineering role. You will be one of the first dedicated AI engineers at Manufex, working directly with the founders to design, build, and ship the agent infrastructure + quoting, assembly, and hardware management from the ground up. There is no existing codebase to inherit, no established team to slot into, and no narrow sprint ticket to execute. You will make architectural decisions that define how the platform works for years. You will own the full AI stack from prompt engineering and agent orchestration to computer vision pipelines and CAD analysis and you will do it in a fast-moving, resource-constrained environment where judgment matters as much as technical skill. If you want to be the person who built the AI behind a category-defining manufacturing platform, this is that opportunity.
What You'll Build
Building AI agents for custom manufacturing workflows across DFM analysis, quoting, supplier operations, compliance, assembly management, and production intelligence. Developing agentic ML systems that ingest CAD files to automatically identify manufacturability risks across CNC machining, sheet metal, assemblies, and 3D printing, while generating instant cost and lead-time estimates based on geometry, tolerances, materials, hardware/BOM requirements, and supplier capacity. Building orchestration and feedback systems that automate RFQs, supplier coordination, compliance verification, inspection validation, and production tracking, with continuous learning loops from real manufacturing outcomes to improve model accuracy, operational efficiency, and decision-making over time.
Responsibilities
Own the full AI/ML stack at Manufex architecture decisions, model selection, infrastructure, and deployment
Ship working agents fast, prioritize functional over perfect, iterate based on real order data
Work directly with the founders on product direction - you will be in the room where decisions about what to build next are made
Build the data infrastructure that makes all agents improvable over time - labeling pipelines, evaluation frameworks, and feedback loops
Define and track the metrics that matter for each agent : accuracy, precision/recall on flagged issues, false positive rates, escalation rates
Partner with operations and quality teams to understand where agents fail and why then fix it
As the team grows, help hire and set the technical bar for the AI/ML function
Requirements
3+ years of hands-on experience building and shipping AI/ML systems in production not research, not prototypes, production systems that real users depend on
Strong foundation in LLM-based agent development prompt engineering, tool use, retrieval-augmented generation, agent orchestration frameworks (LangChain, LlamaIndex, CrewAI, or equivalent)
Hands-on experience with computer vision model fine-tuning, inference optimization, and working with domain-specific visual data; experience with architectures like BLIP-2, SAM, or similar vision-language models is a strong plus
Proficiency in Python and the standard ML stack (PyTorch or JAX, HuggingFace, OpenCV)
Experience with CAD file parsing and geometry analysis (STEP/IGES/STL processing, Open CASCADE, or equivalent) is a significant advantage this is rare and highly valued
Comfort with cloud infrastructure for model deployment AWS, GCP, or Azure; familiarity with containerization (Docker, Kubernetes) and model serving
Strong product instincts you can translate a manufacturing workflow into an agent architecture without being handed a spec
A builder's disposition: you write the code, you debug the pipeline, you talk to the users, you ship
You Might Be a Fit If
You've built agents that do real work, not just answer questions
You've fine-tuned a vision or language model on a small, messy, domain-specific dataset and made it work
You've shipped something in manufacturing, supply chain, logistics, or industrial operations or you're deeply curious about it
You've worked at a startup where the roadmap was a Google Doc and the infrastructure was a single EC2 instance and you made it work anyway
What This Isn't
This is not a research role. We are not writing papers. We are not building demo-ware. We are building agents that manage real orders, review real inspection images, and parse real CAD files for real buyers. If you want to spend six months on a benchmark, this isn't the place. If you want to see the agent you built handling live orders within your first 90 days, read on.
Compensation
Competitive salary benchmarked to US market rates for senior engineers
Meaningful equity as this is a founding role and will be treated as one
Full benefits
Remote-first