Preparation should be balanced between reinforcing your technical foundations and articulating your experience with complex, large-scale systems. Focus on demonstrating how you solve problems from end-to-end.

Technical Depth – We look for candidates who can go beyond using libraries to explain the "why" behind their model choices. Be prepared to discuss the mathematical foundations of your chosen architectures and how they perform under non-ideal conditions.

Engineering Rigor – Building models is only half the battle at Aeva. We evaluate your ability to think about the full deployment lifecycle, including data pipelines, monitoring, and the iterative process of debugging model performance in the field.

Problem-Solving & Adaptability – You will often be presented with ambiguous edge cases. We want to see how you structure your thought process, identify constraints, and propose scalable solutions when faced with incomplete or noisy data.

The interview process at Aeva is designed to be thorough, assessing both your technical mastery and your alignment with our mission to push the boundaries of perception. You can expect a professional, focused experience where you will engage with engineers who are deeply involved in the daily development of our 4D LiDAR technology.

The process typically begins with a recruiter screen to discuss your background and the specific needs of the perception team. Following this, you will move through technical evaluations that emphasize coding ability and domain-specific knowledge. Candidates should expect a rigorous pace that prioritizes practical application over theoretical trivia.

The visual timeline above illustrates the standard progression from initial engagement to technical deep-dives. Use this to pace your study; ensure you are comfortable with your coding fundamentals early on, as these often serve as a gateway to more complex architectural discussions later in the process.

3D Perception & Point Cloud Processing

This is the core of your technical evaluation. We assess your ability to extract meaningful information from LiDAR data.

• Data Representation – Understanding how to represent point clouds (voxels, point-based, etc.).
• Model Training – Strategies for training on long-range or sparse data.
• Evaluation Metrics – Mastering metrics like IoU and mAP for 3D tasks.
• Advanced concepts – Deep knowledge of sensor fusion (e.g., camera-LiDAR integration) and VLM (Vision-Language Model) applications for auto-labeling.

Coding & Software Engineering

Your ability to implement algorithms effectively is critical for production success.

• Efficiency – Writing code that is optimized for performance and memory.
• PyTorch Proficiency – Demonstrating advanced usage of PyTorch for model development.
• System Design – Designing scalable pipelines for data annotation and model iteration.

As a Machine Learning Engineer at Aeva, your primary responsibility is to drive the evolution of our perception stack. You will be responsible for the full lifecycle of 3D perception models, which includes training and evaluating models for object detection, semantic segmentation, and lane detection. You will work closely with hardware and systems engineers to ensure that the models you develop are not only accurate but also performant when running on our silicon photonics-based sensors.

Beyond core model development, you will own the deployment process. This involves diagnosing and resolving issues related to latency and accuracy degradation, as well as building automated pipelines to leverage vision-language models for data annotation. You will be expected to monitor production performance continuously, using real-world data to iterate rapidly and solve the complex edge cases that define the cutting edge of autonomous technology.

A strong candidate for this position combines a solid academic background with practical, demonstrated experience in 3D perception.

• Must-have skills:

  • MS or PhD in CS, Robotics, or a related field.
  • Hands-on experience with 3D object detection on LiDAR point clouds.
  • Proficiency in Python and PyTorch.
  • Experience with large-scale dataset pipelines and annotation workflows.

• Nice-to-have skills:

  • Experience with multi-object tracking or sensor fusion.
  • Familiarity with Vision-Language Models (VLMs) for auto-labeling.
  • Experience deploying models to edge hardware.

Q: How long does the interview process typically take? A: While it can vary, candidates often see the process move from the initial recruiter screen to the technical rounds over the course of a few weeks. We value efficiency and aim to provide timely feedback at each stage.

Q: What is the most important area to study for the coding rounds? A: Focus on your ability to implement core perception algorithms (like IoU calculations) from scratch. The focus is on clarity, correctness, and your ability to explain your logic clearly.

Q: Is this role purely research or production-focused? A: It is a hybrid role. While research is required to stay at the cutting edge, the primary focus is on deploying models that function reliably in the real world.

• Articulate your trade-offs: In system design or architectural discussions, always explain why you chose one approach over another (e.g., speed vs. accuracy).
• Be ready for real-world scenarios: We often present problems based on actual challenges we have faced in the field. Treat the interviewer as a teammate you are collaborating with to solve a problem.
• Understand our tech: Research how our 4D LiDAR technology differs from traditional LiDAR. Understanding the "instant velocity" advantage will give you a significant edge in demonstrating your interest in our specific mission.

The Machine Learning Engineer role at Aeva is an exceptional opportunity to influence the future of autonomous perception. By focusing on your ability to bridge the gap between complex 3D modeling and real-world deployment, you will position yourself as a strong candidate.

We encourage you to revisit the core technical concepts of point cloud processing and sharpen your ability to discuss your past projects with precision and clarity. Your preparation is the foundation of your success; we look forward to seeing how your expertise can help us continue to transform autonomy. Explore additional insights on Dataford to refine your approach, and prepare to bring your best self to the interview.

The compensation data provided covers the broad salary ranges for this role. Candidates should interpret these figures as a reflection of the company's commitment to competitive compensation based on experience, technical skill, and market conditions, with total packages often including significant equity and bonus components.