Staff/Principal Vision Algorithms Engineer

Triton Medical Robotics, a division of Neptune Medical, is an innovative, venture capital funded medical device startup headquartered in Burlingame, CA. We have already commercialized novel medical devices using our technology. Now the team at Triton is building something truly unique: a fully flexible medical robot that enables doctors to reach deep into human anatomy. Our robot will disrupt flexible endoscopy and set a new standard in outcomes for decades to come.

TITLE:  Staff/Principal Vision Algorithms Engineer

POSITION SUMMARY: 

We are seeking a skilled Vision Algorithms Engineer to join our dynamic team. This role will be pivotal in delivering state-of-the-art clinical video quality to doctors using our systems, and computer vision data to our AI engineers.  You will be working with real-time video to improve the accuracy, efficiency, and safety of robotic-assisted medical interventions.  Responsible for architecture, development, and test of algorithms for surgical video applications.

The ideal candidate will act with passion and intensity in a medical device start-up that is fast-paced and collaborative and display the ability to perform to the requirements outlined below. For the right person, this position will be a tremendously exciting and career-defining opportunity, making core contributions to a genuinely new type of medical robotic platform.

ROLES AND RESPONSIBILITIES

• Work with senior leadership and system architects to develop our roadmap for image-processing in endoscopy, delivering subjective video quality for clinicians and high-quality video data for computer vision.
• Design, implement, and improve methods and algorithms for real-time video camera image signal processing (ISP) pipelines, optimizing for latency, reliability, and video image quality in endoscopic environments.
• Build relationship with GPU-ISP technology partner to develop and qualify a new ISP to replace our current pipeline on our NVIDIA Jetson embedded systems, optimized for the unusual features of multicolor-LED-illuminated video inside GI endoluminal passageways.
• Work with clinical engineers and endoscopists to collect subjective image-quality feedback and iterate algorithms and tuning.
• Evaluate potential value of new features like HDR via simulations, post-processing, and benchtop experiments, to prioritize and guide development.
• Create requirements, performance metrics, and evaluation tools for the new ISP.
• Measure performance of new ISP and computer vision algorithms: includes acquiring or developing required tools, making hands-on measurements, and presenting results to partners and internal stakeholders that summarize conclusions at a high level.
• Ensure that all code for new algorithms (internally developed or from our partner) meets the company’s documentation, quality, and compliance standards.
• Integrate available open-source models for rapid prototyping and determination of clinical value

REQUIRED QUALIFICATIONS

• B.S., M.S., or Ph.D. in Computer Science, Vision Science, Physics, or related field.
• 5+ years commercial experience (or 4+ years with M.S., or 1+ years with Ph.D.)
• Passionate about color-quality excellence.
• Expertise in identifying diverse color issues within intricate ISP systems and using the latest advancements in color science, image processing, and machine learning techniques to achieve optimal image quality.
• Good knowledge of state-of-the-art ISP data-flows and processing stages, including auto-exposure, demosaic, noise reduction, white balance, color-matching, contrast-enhancement, and sharpening.
• Enthusiastic to learn CUDA and GPU programming and optimization (experience is a plus).
• Working knowledge of imaging system simulations and evaluation of simulation results.
• Proficient in C++ programming and OpenCV or equivalent.
• Familiar with memory systems and bottlenecks in large-array processing techniques.
• Experience with software documentation, reliability, and compliance in regulated industries or mission-critical applications.

PREFERRED QUALIFICATIONS:

• Expertise and hands-on experience with embedded Linux systems, preferably NVIDIA Jetson
• Experience with HDR
• Knowledgeable in camera sensors, color science, and human perception
• Knowledge of noise models and local tone-mapping
• Experience customizing auto exposure for unusual lighting environments, preferably including developing new algorithms or techniques
• Experience with deep learning frameworks like TensorFlow, PyTorch, and Keras.

Triton Robotics is located in Burlingame, CA, near the Millbrae Caltrain and BART Station.

Our job