AI Agent Software Engineer - Agent Performance Engineering

Great customer support requires human agents and AI in perfect balance, and Assembled is the only unified platform that orchestrates both at scale. Companies like Canva, Etsy, and Robinhood use Assembled to coordinate their entire support operation — in-house agents, BPOs, and AI — in a single operating system. With AI Agents that resolve cases end-to-end, AI Copilot for agent assistance, and AI-powered workforce management that optimizes both human and AI capacity, Assembled helps teams deliver faster, better service while making smarter decisions about how to staff and automate. Backed by $70M from NEA, Emergence Capital, and Stripe, we're building the platform that makes AI and human collaboration actually work.

Our Agent Intelligence team is focused on maximizing AI agent performance and driving volume automation across voice, chat, and email channels. This team is vital for our vision of creating the ultimate omnichannel AI customer support agent and represents one of the fastest growing areas both inside our company and across the industry.

We're currently one of the largest companies deploying AI agents in production at massive scale, working with enterprise clients like Canva, Etsy, and Ramp. The engineering challenges are significant and largely unsolved: building sophisticated evaluation systems, automating knowledge generation, optimizing AI agent accuracy, and creating the infrastructure needed to continuously improve agent intelligence at scale.

As part of the Agent Performance Engineering team, you'll be working on the core systems that make our AI agents smarter, more accurate, and more capable of handling complex customer interactions. You'll be building the evaluation frameworks, knowledge automation systems, and intelligence optimization tools that directly impact our ability to automate customer support at enterprise scale.

• Build foundational evaluation infrastructure: Develop comprehensive evaluation systems from the ground up, including golden dataset creation, automated benchmarking, and model comparison tools. You'll help create the frameworks that enable us to measure and optimize AI agent performance across all communication channels.

• Automate knowledge generation: Design and implement systems that automatically create synthetic guides, documentation, and metadata to improve agent knowledge bases. You'll work on cutting-edge approaches to knowledge extraction and augmentation.

• Optimize AI agent accuracy: Enhance our retrieval systems, implement advanced prompt optimization techniques, and build tools that continuously improve agent responses through automated evaluation and refinement.

• Develop intelligence infrastructure: Architect systems that enable rapid model upgrades, A/B testing of different AI approaches, and scalable evaluation pipelines that support enterprise deployment.

• Drive volume automation: Focus on the north star goal of maximizing automated resolutions across voice, chat, and email by building the intelligence systems that make it possible.

You might be a good fit if you:

• Have 5+ years of experience in software engineering as an individual contributor

• Have experience with AI evaluation systems, data pipelines, or AI model optimization

• Are passionate about building systems that measure and improve AI performance

• Have worked with retrieval systems, knowledge bases, or information extraction

• Enjoy building tools and infrastructure that enable other engineers and AI systems to perform better

• Are highly ambitious and driven, setting high goals for yourself and others

• Put customers first, focusing on solving real problems that impact support quality

• Enjoy fast-paced environments and can quickly adjust when new insights emerge

• Have a bit of a maverick streak that helps you come up with creative solutions

• Have made a noticeable impact on small teams and have solid experience in startups or smaller companies

• Stay humble and open to feedback, value teamwork, and are always ready to learn and grow

• Python/Golang for evaluation systems and data processing

• LLMs and prompt optimization frameworks

• Vector databases and retrieval systems

• AI evaluation and benchmarking tools

• Data pipeline and automation infrastructure

• Generous medical, dental, and vision benefits

• Paid company holidays, sick time, and unlimited time off

• Monthly credits to spend on each: professional development, general wellness, Assembled customers, and commuting

• Paid parental leave

• Hybrid work model with catered lunches everyday (M-F), snacks, and beverages in our SF & NY offices

• 401(k) plan enrollment

We know great candidates don’t always meet every requirement listed in a job description. If the role excites you and you believe you can make an impact at Assembled, we encourage you to apply. We value diverse perspectives and are committed to building an inclusive workplace where everyone feels like they belong and has the opportunity to do their best work. We look forward to hearing from you!

For United States Applicants:
Assembled participates in E-Verify and will provide the federal government with your Form I-9 information to confirm that you are authorized to work in the United States.
For United Kingdom Applicants:
Assembled is required to verify your right to work in the UK and will conduct a Right to Work check prior to employment in accordance with applicable law.