Preparation for the Solutions Architect role at AECOM requires a shift in mindset from pure coding to architectural strategy and operational readiness. You should approach your preparation with the goal of demonstrating how you can deliver value safely and at scale.

Focus on demonstrating strength in these key evaluation criteria:

Enterprise System Design You must demonstrate the ability to design end-to-end distributed systems. Interviewers will evaluate your understanding of backend systems, APIs, and event-driven architectures. You need to show that you can build systems that integrate seamlessly with existing enterprise environments, not just standalone applications.

AI & GenAI Pattern Mastery AECOM is aggressively pursuing AI integration. You will be evaluated on your practical knowledge of AI patterns, specifically Retrieval-Augmented Generation (RAG), agent orchestration, and multi-model workflows. You must be able to discuss the trade-offs between different models, vector databases, and prompting strategies in a production context.

Operational Readiness & Governance In our industry, risk management is critical. You will be assessed on your ability to design for security, data governance, and reliability. You should be ready to explain how you handle versioning, security controls, and system monitoring (observability) for AI solutions.

Stakeholder Partnership You will face questions testing your ability to translate vague business requirements into technical specifications. We look for candidates who can partner with non-technical teams (like Finance or HR), understand their pain points, and propose technical solutions that are cost-effective and scalable.

The interview process for the Solutions Architect role at AECOM is rigorous but structured to assess both your technical depth and your ability to thrive in a collaborative, consulting-driven culture. Generally, the process moves from verifying your background to deep-diving into your architectural capabilities.

Expect an initial screening with a recruiter to discuss your interest in AECOM and your high-level experience with enterprise systems. This is typically followed by a hiring manager interview, which focuses on your specific experience with AI/ML solutions and your alignment with the team's mission. The core of the process usually involves a series of technical deep dives or a panel interview where you will be asked to solution specific business problems, discuss architectural patterns, and demonstrate your knowledge of operationalizing AI.

AECOM values practical, applicable knowledge over theoretical trivia. You should expect discussions to revolve around real-world scenarios—such as automating a complex document review process or optimizing a finance workflow—rather than abstract algorithmic puzzles. The team is looking for evidence that you can build "production-grade" software, not just prototypes.

The timeline above illustrates the typical flow from application to offer. Note that for specialized roles like the AI Solutions Architect, the "Technical Assessment" stage often takes the form of a system design discussion or a case study review rather than a coding test. Use the gaps between stages to research AECOM’s recent digital initiatives and refresh your knowledge on enterprise security standards.

The following areas represent the core pillars of the technical evaluation. You should be prepared to discuss these topics in depth, using examples from your past projects.

AI & Retrieval-Based Architectures

This is the most critical technical area for the Operational AI team. You need to show that you understand how to ground LLMs in enterprise data to prevent hallucinations and ensure relevance.

Be ready to go over:

• RAG Implementation – How to chunk data, select embedding models, and manage vector databases.
• Agentic Workflows – Designing systems where AI agents can take actions (e.g., query a database, send an email) based on reasoning.
• Model Selection – When to use GPT-4 vs. lighter open-source models vs. fine-tuned models based on cost and latency.
• Advanced concepts – Hybrid search (keyword + semantic), re-ranking strategies, and graph-based retrieval.

Example questions or scenarios:

• "Design a system that allows HR employees to query thousands of PDF policy documents and get an accurate, cited answer."
• "How would you handle a scenario where the retrieval system returns outdated information to the LLM?"
• "Explain how you would secure sensitive financial data in a RAG architecture."

Enterprise Distributed Systems

AI does not exist in a vacuum. You must demonstrate that you are a strong backend engineer who understands how to integrate AI components into a broader distributed system.

Be ready to go over:

• API Design – REST vs. GraphQL, rate limiting, and designing contracts for internal services.
• Event-Driven Architecture – Using message queues (Kafka, RabbitMQ) to decouple AI processing from user-facing applications.
• Scalability – Handling spikes in inference requests and managing latency in synchronous vs. asynchronous workflows.

Example questions or scenarios:

• "We have a legacy finance application. How would you architect a modern API layer to expose its data to an AI agent?"
• "Describe a time you had to debug a performance bottleneck in a distributed system."

Operational Excellence (LLMOps)

AECOM requires solutions that are robust and governable. This area tests your maturity as an architect regarding the lifecycle of the software.

Be ready to go over:

• Observability – How to trace a request through a chain of AI agents and log inputs/outputs for auditing.
• Cost Management – Strategies for monitoring token usage and optimizing for cost without sacrificing quality.
• Governance & Security – Data privacy (PII masking), RBAC (Role-Based Access Control), and ensuring AI safety.

Example questions or scenarios:

• "How do you ensure that an AI tool introduced to the IT department doesn't accidentally leak administrative credentials?"
• "Define the metrics you would track to determine if a GenAI solution is 'healthy' in production."