Note: The job is a remote job and is open to candidates in USA. ButterflyMX develops smartphone-based access control solutions that help people open and manage doors and gates. The Principal AI-Native Software Engineer will independently take product initiatives from business requirements through design, implementation, deployment, and operation using Claude Code and coordinated AI agents. The role provides broad ownership across backend engineering, cloud infrastructure, product collaboration, security, testing, and production operations.
Responsibilities
- Work closely with Product to understand business requirements, intended user outcomes, priorities, constraints, and acceptance criteria
- Translate Product's business requirements into clear technical requirements, implementation plans, system behaviors, and engineering tasks
- Identify technical ambiguities, dependencies, risks, edge cases, and tradeoffs that must be resolved before or during implementation
- Ask the questions necessary to ensure the engineering solution addresses the actual product need
- Partner with Design to understand intended user workflows and translate designs into complete, functional product experiences
- Provide Product and Design with clear feedback when requirements or designs introduce unnecessary complexity, technical risk, security concerns, or operational challenges
- Preserve the intent of the product requirements while making sound technical and implementation decisions
- Communicate progress, tradeoffs, risks, and changes in scope clearly throughout development
- Take approved product initiatives from business requirements through technical design, implementation, deployment, and ongoing operation
- Manage several products or major engineering initiatives with limited day-to-day oversight
- Develop complete technical requirements from the business and product requirements supplied by Product
- Make thoughtful architectural, implementation, security, infrastructure, and operational tradeoffs within an established platform direction
- Design, build, and maintain efficient, reusable, reliable, and testable code; identify bottlenecks and bugs, devise solutions, and propose optimizations
- Deliver working software quickly while maintaining high standards for reliability, maintainability, security, and user experience
- Own the technical execution of a product rather than waiting for every implementation detail to be specified
- Continue improving products after launch based on feedback, usage, defects, and changing requirements
- Use Claude Code as a primary development environment and orchestration layer
- Run multiple agents, sub-agents, worktrees, branches, or development sessions in parallel
- Decompose technical requirements into work that can be executed safely and effectively by different agents
- Design reusable Claude Code skills, commands, hooks, context files, agent instructions, and development workflows
- Provide agents with the technical requirements, architectural context, constraints, designs, examples, and verification methods necessary to produce high-quality work
- Build feedback loops in which agents review, test, critique, and improve one another's output
- Use agents to investigate unfamiliar systems, understand legacy code, create implementation plans, write code, generate tests, review changes, prepare pull requests, and assist with releases
- Continuously improve the systems, instructions, and tooling that make your own AI-assisted development more reliable and scalable
- Help automate the engineering portions of the product development lifecycle, including:
- Translating business requirements into technical requirements
- Technical discovery and architecture
- Implementation planning
- Design and UX implementation
- Software development
- Automated testing
- Security analysis
- Code review
- Pull-request creation and iteration
- Infrastructure provisioning
- Deployment
- Monitoring and operational support
- Documentation and maintenance
- Create workflows in which agents can execute substantial portions of the development lifecycle while preserving appropriate human judgment and control
- Establish clear quality gates so that agent-generated work is verified rather than accepted on faith
- Develop methods for measuring the effectiveness, quality, cost, and speed of AI-native engineering workflows
- Reduce the manual coordination and repetitive work required to move from an approved product requirement to deployed software
- Review agent-generated code with the judgment expected of a Principal or Staff engineer
- Reject unnecessary complexity, excessive abstraction, inappropriate frameworks, duplicated logic, and changes that are much larger than the problem requires
- Prefer focused, understandable solutions: for example, an eight-file targeted change rather than an eighty-file rewrite when both accomplish the same result
- Ensure software is appropriately tested, observable, secure, maintainable, and production-ready
- Understand how systems work beneath the abstraction layer rather than treating AI output as a black box
- Debug complex issues across application code, infrastructure, data systems, third-party services, and development tooling
- Know when the AI is wrong, when its proposed approach is risky, and when direct engineering intervention is necessary
- Ensure implementations satisfy the agreed product requirements and acceptance criteria without adding unnecessary scope
- Bring deep, hands-on expertise in our core stack — Ruby on Rails, production TypeScript/Node.js, MongoDB and relational data, and cloud infrastructure across AWS and GCP
- Use Claude Code to move confidently into adjacent or unfamiliar codebases — Android, iOS, Go, and others — developing enough understanding to make sound architectural and implementation decisions without waiting for a specialist
- Work across frontend, backend, mobile, infrastructure, integrations, and internal tooling as needed
- Manage infrastructure and deployment workflows across AWS and GCP
- Improve legacy systems rather than avoiding them or insisting that everything be rewritten
Skills
- Significant experience operating at the Staff, Principal, or equivalent engineering level, with 8+ years of professional software development experience
- A track record of independently delivering complex, production software in partnership with Product and Design teams
- Expert-level experience with Claude Code. This is a mandatory requirement. You must be able to demonstrate substantial product development with Claude Code — not merely autocomplete, isolated code generation, or simple prompting
- Experience coordinating multiple AI agents or development sessions in parallel, with a deep understanding of agent context, skills, sub-agents, tool use, planning, verification, and orchestration
- Production experience with both a dynamic language ecosystem and a strongly-typed one. Production TypeScript/Node.js experience is required
- Experience with relational databases (PostgreSQL, MySQL) and document/NoSQL databases. MongoDB experience is required
- Hands-on experience deploying and operating production software in a distributed, cloud-based environment (e.g., AWS EC2/ECS, Lambda, Cloud Run) within a service-oriented or micro-service architecture, across AWS, GCP, or both
- Hands-on experience applying AI/ML in production. For example integrating LLMs or other models into products or workflows, building retrieval or agentic systems, prompt/eval design, or on-device/edge inference
- Extensive experience developing and maintaining APIs and integrating with third-party APIs
- Strong experience with automated CI/CD pipelines and a deep understanding of testing what, when, and how to test
- Strong ability to translate business and product requirements into technical requirements and implementation plans
- Strong product judgment, including the ability to understand user intent, identify edge cases, and recognize when an implementation does not satisfy the underlying requirement
- Strong understanding of software design, architecture, security, reliability, scalability, and maintainability
- Ability to review AI-generated code critically and distinguish high-quality engineering from superficially functional output
- Comfort working across unfamiliar programming languages, frameworks, and legacy systems, using AI to accelerate that work
- Strong ownership, initiative, and bias toward building
- Excellent written and verbal communication, particularly when expressing technical requirements, constraints, implementation plans, and engineering decisions to both business and technical audiences
- You must have the authorization to work in the US to become an employee
Benefits
- Comprehensive Medical, Dental and Vision plans (ButterflyMX covers 80% of the cost) starting day 1
- 401(k) plan with a match
- 10 paid holidays, 20 vacation days, 5 sick days, 3 floating holidays
- Basic Life and Accidental Death and Dismemberment Insurance (ButterflyMX covers 100% of the cost)
- Short and Long Term Disability (ButterflyMX covers 100% of the cost)
- Paid Family Leave
- Employee Assistance Program
- Quarterly self-care stipends
- Access to optional benefits including pre-tax flexible healthcare spending accounts (FSA and HSA), Dependent Care FSA, and Commuter Benefits, as well as optional Supplemental Life, AD&D, Hospital Indemnity, Legal, Accident, Critical Illness, Pet, and Personal Liability Insurance
Company Overview