Candidates must be United States Citizens, U.S. Nationals, or U.S. Permanent Residents to apply to the NSF I-PERF program.
Candidates must apply to this position through the IPERF portal: https://www.i-perf.org/#/ Schedule: Flexible Monday through Friday. Candidates may be in-person or hybrid. Location: Candidates based in Baltimore, MD or willing to relocate to Baltimore, MD are strongly preferred. Candidates must be able to be in the office an averaged minimum of once per week. What will you do? You will work as a part of our computer vision team to develop the algorithms central to our automated mosquito surveillance products. These products support public health activities in preventing vector borne disease by reducing capacity and capability barriers required to gather vector surveillance data. The most central activity is developing high accuracy fine grained classification algorithms, but various other computer vision tasks such as image processing and object detection are also required to support product function. Previous experience in computer vision, deep learning, and/or image processing through projects, coursework, internships, or employment is expected. Vectech is a startup and responsibilities are dynamic across several exciting, cutting-edge projects. You will have the opportunity to learn new roles, and take on additional responsibilities in our growing team. Responsibilities may include: Develop object detection methods for finding insects or other small arthropods in both standardized and non-standarized image domains Develop classification algorithms for species, sex, and other attributes. This includes improvements on existing and new algorithm development for fine grained problems.Develop conditional GANs to address an imbalanced distribution of attributes in the dataset through data augmentation methods to improve classification algorithm performance. Improving data cleaning and screening human-in-the-loop algorithms, such as ORSAC, to inform the entomology team's label review effortsDevelop methodologies compatible with the available deployment resources Perform scientifically sound testing, with appropriate quality metrics, and when necessary, developing custom metrics or methods for analysis.Identifying and developing new methods using literature review and technical blogs to address specific computer vision and deep learning problemsCollaborating on writing publications on novel work by the teamFollowing repository management protocols for working in collaborationWorking independently on several projects in parallel in a small fast paced teamCollaborating with Computer Vision and Product team members on retrospective analysis to inform model developmentWorking on cloud and local deep learning computers/instances Documenting experimentation and developmentDesired Knowledge* Proficiency in Pytorch, Pandas, OpenCV, FastAIProficiency in Python and with Linux systemsExperience with fine grained classification problems where extremely high accuracy performance is requiredExperience with both classical machine learning algorithms and deep learning algorithmsExperience modifying deep learning architectures for specific development needsExperience collaborating with others on large code repositories Strong communication skills, written and verbalWho is Vectech? Vectech equips professionals with the capabilities of a medical entomologist using AI. Mosquitoes are the deadliest animals in the world, killing over half-a-million people each year. Lyme disease, spread by ticks, is now the most common vector-borne disease in the United States. There are thousands of different species of disease-carrying insects, which carry different diseases and have different behaviors that call for different control measures. With Vectech's products, pest control organizations in public health can quickly generate the information they need to make public health decisions. Our Interview Process Top candidates will be invited to a short phone screening, a short take-home project screening, and a technical interview with our computer vision team. Diversity, equity, and inclusion Vectech is a member of Baltimore Tracks, an organization working to improve diversity in Baltimore tech companies. Women, queer folks, minorities, individuals with disabilities, veterans, and other underrepresented individuals are encouraged to apply and given additional consideration. Candidates must apply to this position through the IPERF portal: https://www.i-perf.org/#/