Algorithms Engineer - Point Cloud Processing

A privately held engineering company is adding a specialist to a small algorithms group working on sensing and measurement.

Capture in the real world is never complete. Whatever the sensor gets, there is always more that it didn't, and the gap is where this role lives. The work is inferring what was never observed, in a way that is physically plausible rather than merely smooth, and then being able to say how far the answer might be off. Accuracy here is a number someone else relies on, so the standard is closer to metrology than to conventional sensing work.

Around that, you would own the processing chain end to end, choose and characterise the sensors feeding it, and take methods from a prototype on your machine to something that runs unattended, without you, in conditions you don't control.

This is a senior individual contributor position rather than a management track. It suits an engineer who finds the underlying problem genuinely interesting and who would rather test an assumption against reality than argue about it.

Sydney based, with flexibility for the right person.

What you would be doing

• Developing methods that work with sparse and partial point cloud data
• Applying physical and mathematical constraints to make results defensible, not just plausible looking
• Selecting, calibrating and characterising sensors, and understanding their failure modes
• Owning the pipeline end to end, covering registration, filtering, error budgeting and output
• Taking methods from prototype through trial into unattended production
• Tracing accuracy problems back through the chain to root cause

What you need

• Strong Python and practical experience with point cloud data
• Fluency in the scientific computing and mathematical stack
• Strong applied mathematics fundamentals
• Experience producing quantitative results from imperfect captures
• Understanding of calibration, registration and error propagation
• Willingness to validate against ground truth in the field, not only in simulation

Please reach out to [email protected] for more info