Senior AI Engineer

The short version

Unburdn helps established companies turn AI from a collection of experiments into working systems that produce measurable business results. We educate teams, identify high-value opportunities, and build production-ready AI workflows and applications that become part of how our clients actually operate.

Client demand has outgrown our current engineering capacity, and we’re hiring another senior technical owner. You’ll own complex client workflows from technical discovery and architecture through implementation, deployment, and production reliability.

You’ll join a small, senior engineering team and work directly with our CTO and our senior engineers. All engineers are expected to own client workflows, contribute to the platform, review one another’s work, and participate in architecture decisions.

Most of your work begins with a specific client problem. Some of it stays client-specific; the best patterns become reusable capabilities in the Unburdn platform. You help decide which is which, and you build both. This is not an ML research role, a prompt-engineering role, or a job where someone hands you finished requirements. It’s senior production software engineering with AI inside it.

What you’ll do

• Own client workflows from technical discovery through production. Translate ambiguous, real-world business processes into systems that are useful, reliable, secure, and maintainable.
• Lead technical scoping. Evaluate feasibility, integrations, data availability, authentication, permissions, security, failure modes, human-review requirements, effort, and delivery risk before commitments are made.
• Build production AI-enabled systems. Develop backend services, workflow orchestration, integrations, data pipelines, document-processing systems, and human-in-the-loop experiences, with application interfaces where needed.
• Work directly with foundation models. Implement structured outputs, tool use, retrieval, classification, extraction, and agentic workflows without treating the model as the entire system.
• Build for the unpredictable parts. Design evaluations, validations, approval steps, fallbacks, exception paths, and operational controls around probabilistic model behavior. Design for the people who’ll actually use the system, not only for the system. An elegant workflow that nobody adopts is a failed workflow.
• Improve the Unburdn platform. Turn what client delivery reveals into reusable connectors, workflow primitives, evaluation tools, approval patterns, observability, security controls, and other shared capabilities. Decide what stays client-specific, what becomes configurable, and what belongs in the core platform.
• Contribute directly to platform architecture. You’ll be empowered to build platform capabilities rather than handing them to a separate team. Foundational changes to the execution model, security boundaries, tenancy, and core architecture are collaborative engineering decisions.
• Own production quality. Establish appropriate testing, logging, monitoring, retries, idempotency, alerting, deployment, and incident-response practices for the systems you build.
• Collaborate as a senior peer. Participate in architecture decisions, review code, challenge assumptions, share context, and help raise the team’s engineering standards.
• Stay current with applied AI. Follow meaningful developments in models, tooling, and infrastructure, but use judgment rather than adopting technology simply because it’s new.

What we mean by “senior”

We don’t require a particular degree, résumé pattern, or exact number of years in the industry. We do require evidence that you’ve already been accountable for production software that real people or businesses depended on.

You can take an incomplete problem, identify what’s missing, make sound architecture decisions, write the code, deploy it, diagnose failures, and communicate tradeoffs without another senior engineer translating each step. You increase the team’s judgment and ownership, not just its coding capacity.

What we’re looking for

• Significant experience building and operating production software systems.
• Strong backend engineering skills, ideally in TypeScript and Node.js. Our environment is centered on TypeScript and Node.js, with Python, relational databases, containers, and cloud infrastructure used where they fit.
• Experience designing APIs, integrations, data models, background jobs, webhooks, queues, and asynchronous workflows.
• Hands-on experience with authentication, OAuth, permissions, secrets management, webhooks, and the operational realities of integrating with third-party systems.
• Experience with relational databases, data transformations, and the messy data that exists inside real companies.
• Experience deploying and operating systems in the cloud, including CI/CD, containers, infrastructure, monitoring, and production debugging.
• Meaningful experience shipping AI-enabled software to real users, plus a practical understanding of foundation-model failure modes and how to design deterministic systems, evaluations, human review, and operational controls around them.
• Clear communication with technical and nontechnical stakeholders, including the willingness to challenge requirements, explain tradeoffs, and say no when a simpler or more reliable solution is the right answer.
• Product and platform judgment. You know when to build a reusable abstraction and when one customer’s edge case should remain an edge case.
• A bias toward ownership. When something you built fails in production, your first instinct is to understand it, fix it, and keep it from recurring.

Especially valuable

• Workflow engines, orchestration systems, or multi-tenant SaaS platforms.
• Enterprise integrations such as Microsoft 365, Google Workspace, Salesforce, Notion, CRMs, ERPs, or document-management systems.
• Document ingestion, extraction, classification, and review workflows.
• AI evaluation, tracing, observability, or regression testing.
• Infrastructure as code, secure code execution, or client-hosted deployments.
• Security-sensitive or regulated environments.
• Solutions engineering, consulting, or other customer-facing work where you stayed hands-on in the code.

What success looks like

First 90 days

• Learn the Unburdn platform, architecture, and existing delivery patterns.
• Become the technical owner of at least one meaningful client workflow and take it from scoping through implementation, testing, and production.
• Deliver at least one reusable platform improvement based on what you learned through client delivery.
• Earn the engineering team’s trust as someone who can carry difficult technical work independently.

First year

• Independently own multiple client systems in production.
• Increase the number of implementations the team can deliver without sacrificing quality, and shorten the path from an approved opportunity to a reliable production workflow.
• Make the platform measurably more durable through better connectors, reusable workflow patterns, evaluation, observability, security, and operational reliability.

Why this might be interesting

• The demand is real. You’ll join a company with active clients and production work ready to be owned, not months of searching for hypothetical use cases.
• The work gets used. We measure success by what reaches production and how it changes the way a client operates.
• You’ll have strong technical peers. You’ll have people to debate architecture with, without layers separating you from the decisions or the customer problem.
• You’ll shape the platform through evidence. The platform is evolving alongside real implementations, and you’ll have direct influence over what gets built and why.
• You’ll see the whole problem. Understand the business process, talk to the people who perform it, design the system, build it, and see whether it created value.
• We care about practical AI, not hype. We use the latest technology when it improves the result and conventional software when that’s the better answer.

This probably isn’t the right role if

• You want to focus on model research or training.
• You prefer prototypes to production ownership.
• You want a pure platform role with no exposure to customers or business processes.
• You need detailed specifications before you can start.
• You’d rather hand architecture to another team than build the system yourself.
• You adopt new frameworks faster than you evaluate whether they’re reliable.
• You’re looking for a junior or narrowly scoped implementation role.

How to apply

Send your résumé or LinkedIn profile and briefly tell us about one production AI-enabled system you personally owned. What problem did it solve? What did you build? Which important architecture decisions did you make? How did you evaluate its behavior, handle failure cases, and decide where human review was required? What happened once real users began using it?

You may also include code, a product, a technical write-up, an architecture explanation, or another example of your work. We understand that much of your work may be private, so this is optional.

Unburdn is an equal opportunity employer. We evaluate candidates based on their ability to do the work and contribute to the team, and we welcome applicants with nontraditional backgrounds.