Xihelm is developing the first viable automated picking robot for glasshouse vegetables. You'll be helping to make that a reality - focused on both machine learning and perception (in 3D).Responsibilities
- Novel deep neural network design
- Potentially developing a new design and iterating on current designs
- Implementation of network using current frameworks (PyTorch and TensorFlow)
- Training using local and cloud compute
- Deployment and monitoring of continually retrained models
Understanding of existing segmentation techniques (such as detectron) Perception
- Competent in the manipulation of point clouds
- Oversight on all automated robot perception
- Marker detection based visual techniques
- 3D, point cloud-only calibration
- Refinement techniques and benchmarking
- Designing lab & field experiments that prove efficacy of your proposed improvements - "hypothesis-driven development"
- PhD in a technical field (involving machine learning, computer science, robotics, physics, electrical engineering or equivalent), ..or recent masters in very related field
- Working knowledge of PyTorch or similar frameworks
- Experience with
- implementing numerical methods and data visualisation
- SQL/data retrieval
- Build processes
- AWS/Google cloud services
Having a passion for AI and roboticsShould be willing to integrate models into our existing ROS stackA cover letter explaining why you are a fit for this specific role at Xihelm - if you're lazy, or just applying for every job you're almost certainly wasting your time.Notes
This job will need occasional UK and international travel including overnight, typically less than 2 days a month.
We hire on a rolling basis, and interviews may be cancelled if we find a suitable candidate.
Location London-based, but you'll be able to work from home typically. This should allow easy access to our robotics lab in London (SE16).
UK visa sponsorship possible.
Hours this job requires keeping UK working hours (9am to 5pm, Monday-Friday). It may involve some irregular (or on-call) out of hours work in future.
Salary £40,000 - £80,000 annually, depending on experience.
Posting date 2/Sep/21
Job expiry date 1/Oct/21
Hiring process (can be done in
- Do-at-home task (2hrs) with a 1 week hard submission time (please do not ask for an extension - we won't review).
- Screening call (20mins) to discuss role, ask basic questions
- 2-3 x 30min tech (including live coding)
- 1x 1-2hr final stage CV/resume deep dive
- Reference calls
- Offer subject to contract etc.
- Strong experience with Python (python2/3)
- Designed/implemented/evaluated Computer Vision neural network architectures
- Mathematical background of Linear Algebra and Calculus at least at undergraduate level
- Linux (preferably Ubuntu)
- 25 working days holiday (plus bank holidays)
- Company pension contributions
- Private health, dental and life insurance
- Meaningful, EMI-compliant share options - so when we make it big - you do too!
- Regular team offsites/fun events (easier in a post-Covid world)
- Learning & development budget