Job Description

Customer experience is at the heart of all we do at Amazon, as we strive to be Earth's most customer-centric company. Amazon Customer Service (CS) is at the forefront of customer experience, working relentlessly to provide customers with a convenient, high quality, and delightful experience - across all our CS channels, in every interaction.

Within CS, Human in the Loop (HITL) engineering brings human judgment into AI-handled conversations in real time, across every use case, channel, and marketplace where automation reaches its limit. We build the systems that predict when a bot is about to fail. We summarize live conversation context so a Customer Service Associate can decide in seconds, and we execute that decision back into the conversation while the customer is still in it, with no transfer and nothing to re-explain. Every intervention also becomes training data for the next contact like it, so the case that needed a human this time is one automation can handle on its own the next. That is why HITL compounds instead of plateauing.

We are looking for a Software Development Engineer to help build that platform and shape the architecture behind it. Today HITL runs tens of thousands of customer contacts every week across four use cases in the US, UK, and Canada. What stands between us and ten times that is the architecture itself. Core logic is spread across several services, each new use case takes about six weeks to launch, and tight coupling to our current bot stack blocks entire product lines where human judgment would matter most. You will be a significant contributor to fixing that. You will help design and build HITL as a standalone platform, shape the contracts that let any bot in CS request human judgment, and make the real-time path fast and reliable enough that a customer sitting in a live conversation never knows a human was involved.

This is a high-ambiguity, high-ownership role. You will own components end to end, from the design doc through launch, and you will help make the calls on latency budgets, failure modes, and what the platform guarantees to the teams building on it. The problems are hard ones:

• Orchestration across multiple customer service systems with no single owner
• LLM inference inside a conversation a customer is waiting in
• Triggering logic that has to be right at the moment it fires, not just in aggregate
• A keep-alive mechanism that will not survive our next scale-up
• Feedback pipelines that turn every associate decision into training data worth learning from.

You will work with applied scientists to move models from notebook to production path. You will partner with the teams who own the systems HITL depends on, and you will contribute to the engineering standards of the team through code and design review and by mentoring the engineers around you.

The person we are looking for likes problems where the right design is not obvious yet, cares about operational excellence because customers feel every regression, and wants to build AI systems that get measurably better every week because of the decisions they made.

Key job responsibilities

• Own the design and delivery of major HITL components, from technical design through production launch and ongoing operation.
• Contribute to the effort to decouple HITL into a standalone platform, helping define the service contracts and extension model that let new bots and use cases onboard in days rather than weeks.
• Make and defend architectural decisions on latency, reliability, and failure handling for components that sit inside live customer conversations.
• Partner with applied scientists to productionize ML and generative AI components, including context summarization and proactive questioning, and own the serving path, evaluation harness, and fallback behavior.
• Drive operational excellence. Own the metrics, alarms, and on-call for your systems, run root cause analysis on the failures you find, and go after whole classes of failure rather than one instance at a time.
• Participate actively in design and code review, including reviews of more senior engineers, and mentor engineers on the team.
• Partner with product managers, science, and dependent teams to turn ambiguous customer experience problems into scoped technical plans, and push back when requirements do not survive contact with the architecture.
• Contribute to roadmap and planning, including flagging the technical investments the team should make before they turn into blockers.

A day in the life

Some days are heads-down implementation on a service in the live contact path. Others are a design review where you are defending a contract decision that partner teams will build against, or a production investigation where a latency regression is something customers are feeling right now. You will balance the current use-case launch against the platform work that makes the next ten cheap, and you will help decide where that line sits. You will spend real time with applied scientists making models production-viable, and real time with the engineers around you making their designs better.

About the team

Customer Service Technologies, part of Amazon Customer Service (CS), builds the platforms and products that power Amazon's world-class customer service experiences. Our team of scientists and engineers develops and deploys LLM-based Conversational AI systems that help customers solve their issues and get their questions answered efficiently, as well as associate-facing products that support our customer service workforce.

Basic Qualifications

• 3+ years of non-internship professional software development experience
• 2+ years of non-internship design or architecture (design patterns, reliability and scaling) of new and existing systems experience
• 2+ years of designing and developing large-scale, multi-tiered, multi-threaded, embedded or distributed software applications, tools, systems, and services using: C#, C++, Java, or Perl experience
• 2+ years of Object Oriented Design experience
• Bachelor's degree or foreign equivalent in Computer Science, Engineering, Mathematics, or a related field
• Experience programming with at least one software programming language

Preferred Qualifications

• 3+ years of full software development life cycle, including coding standards, code reviews, source control management, build processes, testing, and operations experience
• Knowledge of machine learning model architecture and inference

Amazon is an equal opportunity employer and does not discriminate on the basis of protected veteran status, disability, or other legally protected status.

Our inclusive culture empowers Amazonians to deliver the best results for our customers. If you have a disability and need a workplace accommodation or adjustment during the application and hiring process, including support for the interview or onboarding process, please visit https://amazon.jobs/content/en/how-we-hire/accommodations for more information. If the country/region you’re applying in isn’t listed, please contact your Recruiting Partner.

The base salary range for this position is listed below. Your Amazon package will include sign-on payments and restricted stock units (RSUs). Final compensation will be determined based on factors including experience, qualifications, and location. Amazon also offers comprehensive benefits including health insurance (medical, dental, vision, prescription, Basic Life & AD&D insurance and option for Supplemental life plans, EAP, Mental Health Support, Medical Advice Line, Flexible Spending Accounts, Adoption and Surrogacy Reimbursement coverage), 401(k) matching, paid time off, and parental leave. Learn more about our benefits at https://amazon.jobs/en/benefits.

USA, WA, Seattle - 143,700.00 - 194,400.00 USD annually