Senior Back End AI Cloud Engineer

About Altus Nova
Altus Nova combines strategy consulting, product management, and software development into a single service. We work closely with client executives and front-line teams to understand business processes, define practical solutions, and deliver value through phased roadmaps. Our work spans healthcare, financial services, ecommerce, logistics, and infrastructure. We remain accountable for the systems we build and often operate and support them after delivery.

The role
You will design, build, and evolve cloud-native back-end systems, using AI to accelerate disciplined engineering. You will independently apply agreed architectural and engineering practices, work directly with clients and teammates to clarify requirements, and own your work through validation and support. Strong engineering fundamentals and judgment are essential to directing AI effectively.

What you will do
Analyze business workflows and requirements with clients and the product team. Surface contradictions, missing requirements, risks, and impractical assumptions before directing implementation.
Develop technical designs, interface contracts, implementation plans, and acceptance criteria. Maintain shared specifications and decision records as requirements and solutions evolve.
Design and build modular back-end capabilities and integrations with clear boundaries. Evaluate changes for compatibility, dependencies, security, performance, and supportability.
Direct agentic development tools through bounded tasks and explicit instructions. Review generated changes and retain ownership of code quality and the overall solution.
Deliver in small, testable iterations. Use focused research spikes to resolve technical uncertainty, and contribute to estimates and transparent delivery planning.
Build automated validation and diagnostic capabilities. Work with QA and DevOps on release readiness, secure deployment, production support, and learning from incidents.

Required experience and engineering foundations
5+ years of professional back-end engineering experience. The 12-month AI experience requirement below is part of this total.
Strong computer science and software engineering fundamentals, grounded in substantial hands-on back-end development with a production stack such as C#/.NET, Python, or Node.js/TypeScript.
Hands-on production experience with at least one of AWS, Azure, or GCP, including serverless compute, messaging, data storage, identity, observability, and deployment.
Sound understanding of API design, service and module boundaries, SQL and NoSQL data modeling, distributed systems, concurrency, and event-driven processing.
Practical experience with automated testing, code review, source control, CI/CD, infrastructure as code, application security, and diagnosing production problems.
Ability to read, critique, debug, and modify code and resolve technical issues independently of AI. You remain accountable for designs, implementations, and outcomes regardless of how they are produced.

Required experience with design time AI
Design-time AI means using AI during analysis, design, implementation, testing, and maintenance. We require practical proficiency in rigorous engineering across repeated changes to complex systems.
At least 12 months of sustained professional AI-assisted software development within your overall engineering experience, including hands-on use of agentic coding tools such as Claude Code, Codex, or comparable tooling.
Experience delivering production software with AI and maintaining it through multiple subsequent iterations. You can explain your specifications, architectural choices, change controls, validation, and lessons from failures. Vibe coding alone does not meet this requirement.
Ability to apply disciplined, specification-driven development: clarify ambiguity with the appropriate owners, expose assumptions, define technical designs and acceptance criteria, and keep specifications, contracts, decisions, and implementation aligned.
Effective delegation to AI: decompose work into bounded tasks, provide relevant context and constraints, anticipate misinterpretation, define acceptance criteria, and critically review outputs. Use adversarial review to challenge assumptions and surface gaps.
Ability to discuss choosing models for analysis, implementation, and review based on capability, reliability, latency, and cost, and managing context and token consumption in development workflows.
Understanding of how architectural boundaries, explicit interfaces, and contracts contain change. Explain how you would prevent or detect unintended rewrites, assess downstream effects, and maintain a complex application across iterations.
Ability to discuss validation controls, their limitations, and when particular controls are mandatory or implementation choices. Recognize that AI-generated code and tests can share mistaken assumptions and explain how you would challenge them.
Ability to describe how specific tools and workflows preserve shared project context and enable another engineer to continue the work. Willingness to learn and work with Altus Nova Assistant (ANA), Altus Team AI, and the AI technologies and frameworks we develop.

Required runtime AI understanding
Runtime AI means AI capabilities operating within deployed solutions. You must meaningfully discuss the concepts below and outline a basic production workflow. Practical runtime AI implementation experience is a strong advantage.
Explain model limitations, context management, retrieval, structured outputs, tool use, and cloud-based agentic workflows, with the ability and commitment to mature into production implementation.
Distinguish functionality best implemented with deterministic code from work suited to LLM reasoning. Discuss how the approaches can be combined within reliable systems.
Discuss workload routing among models based on capability, reliability, latency, and cost. Explain context and token management and guardrails for budgets, retries, and agent execution limits.
Discuss safeguards that reduce and detect hallucinations, including grounding, output validation, evaluation, and appropriate abstention or escalation when evidence is insufficient.
Distinguish generating a recommendation from executing an action. Discuss least-privilege access, human review and approval, rejection paths, failure handling, observability, and recovery.
Discuss protecting client information in development tools and runtime systems, including secrets management, prompt injection, untrusted retrieved content, and restrictions on agent access to tools and systems.

Communication and collaboration
- Strong spoken and written English is required. You must explain complex concepts in plain language, listen carefully, ask precise questions, and communicate effectively with client executives, subject-matter experts, and engineering teammates. Constructively challenge flawed specifications and work with their owners to resolve them.
- You must help teammates succeed, surface blockers early, and provide clear handoffs. Effective delegation to AI requires instructions others can understand and outputs you can explain and defend.
- You will work with Product Managers, Business Analysts, architects, designers, front-end engineers, QA, and DevOps. We value constructive debate, clear decisions, and helping teammates succeed. Engineers engage directly with clients, participate in planning and estimation, demonstrate their work, document decisions, and track time accurately. Production reliability is a shared priority.

Preferred experience
C#/.NET development and experience adapting across back-end languages and frameworks.
Production AI workflows, including model routing, retrieval, evaluations, human approval, and cost controls.
Legacy modernization, third-party integrations, or business-critical systems in complex client environments.

How we assess candidates
Our process includes a collaborative, real-time whiteboard session where you analyze unfamiliar problems, clarify ambiguity, and explain your approach with the interviewers. Candidates who pass complete a practical task using AI and walk the team through the resulting specifications, implementation, and validation. You must demonstrate independent reasoning, understand the outputs, and be able to take responsibility for the outcome.