Software Engineer III

Meta Platforms • Sunnyvale, CA

1d

About The Position

The Software Engineer III will operate and improve camera and sensor simulation pipelines to generate synthetic datasets on demand. The role delivers training and evaluation data to camera architects and machine-learning teams under tight timelines and supports ML training and evaluation using generated data. This includes running training jobs, producing evaluation metrics, and iterating on dataset composition based on model performance.

The position configures simulation scenes and sensor parameters to meet customer and architect specifications, validates and quality-checks generated data before hand-off, and maintains and improves simulation tooling and dataset-generation scripts. The role is based in Sunnyvale, California, for a duration of 12+ months.

Requirements

• Bachelor's degree in Computer Science or a related field.
• At least 5 years of software engineering experience, focused on ML infrastructure design.
• 5 years of experience in Python and C.
• Knowledge of camera and sensor simulation fundamentals, including camera intrinsic and extrinsic setup and calibration.
• Knowledge of sensor modeling, including noise sources, conversion gain, quantum efficiency, and read and shot noise.
• Basic understanding of ISP and image formation.
• Experience running ML training and evaluation workflows; PyTorch is preferred.
• Ability to interpret model metrics and translate them into dataset changes.
• Comfort with large-scale dataset generation, storage, and versioning.
• Familiarity with Linux development environments and source control.
• Experience building ML models and pipelines.
• Understanding of how camera simulations work.
• Previous AI experience developing ML architectures.

Nice To Haves

• Experience with rendering engines such as Blender, Unreal, or Unity, or with in-house simulation frameworks.
• Optical or imaging background, including radiometry, PSF, or MTF.
• Prior experience with camera or ISP tuning or perception model development.

Responsibilities

• Run and iterate on camera and sensor simulation pipelines to generate synthetic datasets on demand.
• Deliver training and evaluation data to camera architects and ML teams under tight timelines.
• Support ML algorithm training and evaluation using generated synthetic data by running training jobs, producing evaluation metrics, and adjusting dataset composition based on model performance.
• Configure simulation scenes and sensor parameters to match customer and architect specifications.
• Validate and quality-check generated data before hand-off.
• Maintain and improve simulation tooling and dataset-generation scripts.

What This Job Offers

• Career Level: Mid Level