Senior Software Engineer – Computer Vision Optimization
Metalenz is at the forefront of driving innovation in optics with metasurface technology, providing solutions that redefine the possibilities of mobile imaging and sensing. Metalenz is the first company to bring metasurfaces to mass markets, with millions of its meta-optics already integrated in consumer devices, combining the functionality of three or four complex lenses and components into a single flat device, mass-produced in existing semiconductor foundries.
The Company’s latest product, Polar ID, is a groundbreaking imaging sensor solution that can see what no other sensor can see, enabling secure and affordable face unlock for all smartphones. This system-level solution harnesses the unique capabilities of metasurface optics together with machine vision algorithms to bring a new form of sensing to mobile devices for the first time.
Metalenz is built on foundational research from Harvard University, and backed by venture arms of Intel, 3M, TDK, and others.
The Position
Metalenz is developing next-generation sensing systems for consumer electronics with metasurface technology. As a Software Engineer, you will develop computer vision and image processing algorithms to work with our engineered meta-optic camera systems. A successful candidate will be able to implement existing and new computer vision algorithms for 3D scene reconstruction, segmentation, and biometric tasks.
What you’ll do
- Optimize machine learning model implementations and inference times
- Work with TFLite, SNPE, or other embedded deep learning libraries
- Quantize and tune model architecture, speed, and performance while maintaining task performance
- Participate in rigorous benchmarking and analysis across model classes
- Deploy to mobile hardware running Android
- Conceive, implement, and optimize deep learning and computer vision algorithms using our unique meta-optic capabilities to bring truly secure authentication to mass market
- Work in facial identification, presentation attack detection, and scene synthesis
- Implement your algorithms on our unique camera hardware
- Assist software team in efficient model integration for application and service hosting the models by writing C++/JNI/Kotlin
- Collaborate by writing documentation, engaging in code review, and commenting on design documents
What you’ll need
- MS or PhD in Computer Science, Electrical Engineering, or relevant field and/or demonstrated excellence in relevant technical domains through open source contributions or publications
- 5+ years’ experience building computer vision solutions and deploying deep learning models on mobile
- Broad understanding of methods and tooling, from traditional (FastCV) to deep learning approaches (i.e. PyTorch, Jax, Tensorflow) and their deployment to embedded systems (Android, C++)
- Knowledge of cameras, image processing, and computer vision techniques
- Familiarity with software engineering best practices (version control, data provenance, code profiling, and unit tests)
- Knowledge of building and deploying machine learning train and test pipelines a plus
- Proficiency in Python, C++, and JNI
What we do
- Work with consumer electronic OEMs to commercialize meta-optics
- Design next-generation optical systems for emerging consumer applications like 3D sensing
What We Offer
- We provide competitive compensation
- Pre-IPO stock options
- 100%, company paid premiums for employees and their dependents, such as medical, dental, and vision
- Employer Health Savings Account (HSA) contributions
- Employer matching 401k
- Pre-tax commuter plans
- Casual dress code
- 18 days of PTO and 12 paid company holidays
- 10 sick days per year on a calendar basis
Metalenz is an equal-opportunity employer. The Company is committed to providing an open and inclusive environment for all of its employees. The Company prohibits any discrimination on the basis of gender, gender identity, pregnancy, race, color, religion, sexual orientation, national origin, ancestry, age, disability, veteran status, criminal record, or genetic information.
#J-18808-Ljbffr
