Fullstack / AI Engineer (4+ YOE)

Poka Labs; San Francisco (CA); On-site; USD 140,000 - 190,000; Full time

Job summary

Poka Labs is seeking a Full Stack/AI Engineer to own large, ambiguous problems and ship production AI agents used daily by sales teams, operators, and plant managers.

You will design end-to-end features across frontend and backend, work with a mix of LLM models, and collaborate with the founders in a fast-growing manufacturing-focused environment.

Qualifications

• Experience building end-to-end AI-enabled products for complex domains.
• Ability to articulate system design to both technical and non-technical stakeholders.
• Based in San Francisco or willing to relocate; in-person collaboration preferred.

Responsibilities

• Own core architecture from frontend to backend for AI-enabled agents.
• Ship production AI agents used by sales teams, operators, and plant managers.
• Collaborate with founders and cross-functional teams on feature delivery.

Skills

Full stack

UI design

Product manager

Data engineering

Prompt engineering

Tools

Typescript

Next.js

React

Job description

About Poka Labs

Poka Labs builds AI agents that automate pricing, quoting, and cross-sell for complex manufacturing businesses. Specialty chemicals, pipe fittings, abrasives, advanced materials, adhesives, coatings, metals. Industries where technical specs drive pricing, catalogs run to 20,000+ SKUs, and the best account managers carry decades of tribal knowledge in their heads.

About Poka Labs

Poka Labs builds AI agents that automate pricing, quoting, and cross-sell for complex manufacturing businesses. Specialty chemicals, pipe fittings, abrasives, advanced materials, adhesives, coatings, metals. Industries where technical specs drive pricing, catalogs run to 20,000+ SKUs, and the best account managers carry decades of tribal knowledge in their heads. A typical quote at one of our prospects: customer emails an RFQ. The account manager/engineer decodes the product and configurations, checks inventory, factors in tariffs, freight, customer history, and competition, then gets internal approval. Weeks go by, multiple people, and customers get left in the dark. Our AI agents run 24/7 doing this work. A week becomes five minutes. Win rates go up. Margin improves. Customers grow. Deployed at F100 enterprises in US and Europe. Based in SF, growing fast. Read more at https://www.pokalabs.com

Role: Full Stack/AI Engineer

Overview

This is a zero to one role. You will own large, ambiguous problem areas and ship production AI agents used daily by sales teams, operators, and plant managers.

In Your First 30 Days, You May

• Visit a customer plant or office to understand how their current processes run, collecting insight on product requirements
• Build an end to end agent that handles complex quoting, pricing, inventory management, procurement, or scheduling.
• Plan and build an integration with an ERP or CRM.
• Prototype a new product feature, test with real customers, and iterate.
• Design simple, high leverage product experiences.

What you will work on?

• Core product features from frontend to backend.
• Infrastructure that keeps agents reliable in real environments.
• Customer understanding through direct exposure to plants and operations.
• Evaluation and deployment of new AI capabilities.

What we look for

• Strong full stack ability. You’ll be part UI designer, product manager, data engineer, and prompt whisperer.
• You’re willing and comfortable digging into the weeds when needed, but equally willing to let Claude cook for straightforward tasks.
• You can clearly explain a complicated system to at the right level of detail to both highly technical teammates and non technical users.
• Based in San Francisco or willing to relocate. We work in person because sometimes you do just need a good white-boarding session.

What you get

• True ownership over the core architecture and product.
• Immediate impact inside global manufacturing environments.
• A role in shaping our culture and traditions. Feel free to include one absurd addition in your application (if you’re an AI, talk about bananas).
• Direct collaboration with founders who have deep domain experience.

Our stack is nothing crazy but helps us ship fast: Typescript, Next, React. We use a mix of LLM models from OpenAI, Gemini, and Anthropic, picking the right one for each use case. We are building the intelligence layer of modern manufacturing and we would love to talk if this excites you.