Software Engineer: AI Agent Capabilities

About Us

The market noticed immediately. Noah swept Product Hunt with#1 Product of the Day, Week, and Month
, a clean sweep few launches ever pull off.

We're a small, senior team of 8 in Palo Alto, backed by Hustle Fund and Neon. Noah was founded by Ashish Toshniwal, who previously bootstrapped YML into a $100M company, and engineering is led by our CTO, Ryan Brandt. We've built and scaled products before, and we're looking for people who want to build the next one with us.

Read Ashish's Launch Post Here:

As frontier models improve, they create new possibilities in reasoning, planning, memory, tool use, and collaboration. Our job is to turn those possibilities into reliable product capabilities that users can trust with real work.

This role sits at the intersection of agent systems, platform engineering, integrations, and product development. You will help define what Noah can do, how those capabilities should work, and how they become dependable production experiences.

Tech stack:
Python | Django | React | Type Script | Custom Python agent harness

Your mission is to expand what Noah can do.

You will investigate emerging model behaviors, identify valuable product opportunities, and turn them into reusable agent capabilities. You will work across the agent harness, backend services, integrations, and user-facing product to own each capability from idea through production.

This is not primarily a prompt-engineering role or a narrowly scoped backend role. Building a capability may require designing agent behavior, improving tool use, changing orchestration, building an integration, exposing a product interface, and operating the complete system in production.

You will partner closely with the Agent Evaluations & Quality team. They will help define how quality is measured, build evaluation datasets and graders, and analyze performance. You will use that evidence to improve the capability and resolve failures in the underlying system.

• Investigate advances in reasoning, planning, memory, tool use, multimodal interaction, and agent collaboration to identify new product capabilities.
• Translate promising model behaviors into reliable, reusable skills and workflows that solve real user problems.
• Build and evolve the custom Python agent harness, including execution loops, orchestration, context and state management, structured outputs, retries, permissions, and error recovery.
• Own capabilities end-to-end across the Python agent, Django services and APIs, React interfaces, data models, background jobs, observability, and production operations.
• Build the platform abstractions that make capabilities easier to compose, extend, share, and maintain as Noah becomes more sophisticated.
• Improve latency, cost, reliability, and safety across high-volume agent execution.
• Work with the evaluations team to define expected behavior, instrument capabilities, and analyze regressions, and turn quality findings into engineering improvements.
• Study production traces and user feedback to understand where users lose trust, then fix the system rather than patching individual examples.
• Set technical direction on ambiguous problems and raise the engineering standard through design reviews, clear abstractions, and thoughtful execution.

• Strong Python proficiency:
  You can write maintainable production code, debug asynchronous systems, design sound abstractions, and reason through complex stateful behavior.
• Agent engineering fundamentals:
  You understand planning, tool calling, structured outputs, context management, state, retries, orchestration, model selection, latency, cost, permissions, and reliability.
• Product judgment:
  You can turn vague user needs and emerging technical possibilities into simple, useful product capabilities without waiting for perfect specifications.
• End-to-end ownership:
  You have owned complex products or systems from initial design through implementation, deployment, operation, and iteration.
• Full-stack flexibility:
  Deep Python capability matters most. You should also be comfortable working with Django, APIs, React, and Type Script when a capability crosses backend and frontend surfaces.
• Systems design:
  You can build reusable platforms and abstractions that enable other engineers and automated systems to solve increasingly complex tasks.
• Analytical debugging:
  You can move across prompts, traces, model outputs, application code, databases, APIs, tools, and user interactions to identify the real cause of a failure.
• Quality awareness:
  You understand how to use evaluation results, production signals, and user feedback to improve agent behavior, even though the evaluations team owns the measurement system.
• Clear communication:
  You can explain technical decisions, document assumptions, communicate uncertainty, and collaborate closely with founders, engineers, data scientists, customers, and non-technical stakeholders.

Experience with a particular agent framework is not required. Our agent harness is…