To excel in your interviews, you must prepare deeply across several distinct technical and behavioral domains. Here is a breakdown of what Bentley Systems typically evaluates.
Software Engineering and Algorithms
As an AI Engineer, your code must integrate seamlessly into Bentley’s broader enterprise software ecosystem. This area tests your ability to write efficient, bug-free code under pressure. Interviewers are looking for strong fundamentals in time and space complexity, edge-case handling, and code readability.
Be ready to go over:
- Data Structures – Arrays, hash maps, trees, and graphs. Graph algorithms are particularly relevant given Bentley’s focus on spatial and network data.
- Optimization – Identifying bottlenecks in your code and refactoring for better performance.
- Object-Oriented Design – Structuring your code logically, which is critical when interfacing with complex CAD or digital twin APIs.
- Advanced concepts (less common) – 3D computational geometry basics, spatial indexing (like KD-trees or Octrees).
Example questions or scenarios:
- "Write a function to find the shortest path between two nodes in a highly connected graph representing a utility network."
- "Given a stream of sensor data from an IoT device on a bridge, how would you efficiently compute the moving average over a sliding window?"
- "Design a class structure to parse and manipulate a large JSON file containing 3D bounding box coordinates."
Machine Learning and Applied AI
This evaluation area tests your depth of knowledge in building, training, and evaluating AI models. Bentley works heavily with computer vision, predictive maintenance, and increasingly, generative AI. You need to prove you can move beyond calling APIs and truly understand model behavior.
Be ready to go over:
- Computer Vision – Object detection, image segmentation, and handling 3D point cloud data (LiDAR).
- Model Evaluation – Choosing the right metrics (Precision, Recall, F1, IoU) and diagnosing issues like overfitting or data leakage.
- Generative AI & LLMs – Fine-tuning strategies, RAG (Retrieval-Augmented Generation) architectures, and prompt engineering for enterprise data.
- Advanced concepts (less common) – Physics-informed neural networks (PINNs) or time-series forecasting for structural health monitoring.
Example questions or scenarios:
- "How would you design a computer vision model to detect structural cracks in high-resolution drone imagery?"
- "Explain the tradeoffs between using a pre-trained LLM via API versus fine-tuning a smaller open-source model for querying internal engineering documents."
- "If your object detection model performs well on training data but fails on images taken in low-light conditions, how do you troubleshoot and resolve this?"
AI System Design and MLOps
For Senior Software Engineer and Applied AI Solution Engineer roles, system design is a critical differentiator. You must demonstrate how to take a model from a local environment and deploy it reliably in the cloud or at the edge.
Be ready to go over:
- Model Serving – REST APIs, gRPC, batch processing vs. real-time inference.
- Data Pipelines – Handling massive datasets, feature stores, and ETL processes.
- Monitoring and Maintenance – Detecting model drift, managing retries, and handling system failures gracefully.
- Advanced concepts (less common) – Distributed training architectures, optimizing model inference latency using TensorRT or ONNX.
Example questions or scenarios:
- "Design an end-to-end system that ingests daily satellite imagery, runs an anomaly detection model, and alerts users of potential land subsidence."
- "How would you architect a RAG pipeline that needs to securely index and query millions of proprietary PDF engineering schematics?"
- "Walk me through how you would set up CI/CD for a machine learning model updating weekly."
Behavioral and Culture Fit
Bentley Systems values engineers who are collaborative, adaptable, and focused on user impact. This area evaluates your past experiences, your ability to handle conflict, and your communication style.
Be ready to go over:
- Navigating Ambiguity – Taking vague product requirements and turning them into concrete technical plans.
- Cross-Functional Communication – Explaining complex AI concepts to non-technical stakeholders or domain experts (like civil engineers).
- Ownership and Impact – Discussing projects where you took the lead, made mistakes, learned, and delivered value.
Example questions or scenarios:
- "Tell me about a time you had to compromise on model accuracy to meet a strict latency or deployment constraint."
- "Describe a situation where you disagreed with a product manager about the technical direction of an AI feature. How did you resolve it?"
- "Walk me through a project that failed. What went wrong, and what did you learn?"
