Full Job Description
Computer Vision / ML Engineer, Gaia AI
We are currently building our team and hiring for a full-time, in-person Computer Vision Engineer, with the ambition to help shape and drive our development with implementing our computer vision pipeline. This is a great opportunity for someone who wants to apply their skills towards fighting climate change, wants to experience the high-pace, high-ownership setting of an early startup, and can start onboarding immediately or within a few weeks. This is an in-person role in Boston, MA.
Climate change is one of the most meaningful challenges of our generation, and every possible solution put forward by the IPCC requires scaled carbon sequestration. Forests are critical for this reason, but a lack of trust is holding back the market for forest-based carbon credits. We are an MIT climatetech startup applying technology we are used to building in the autonomous vehicles space to solve this problem, utilizing cutting edge sensors and perception AI to verify forest-based carbon credits with great accuracy and confidence.
At Gaia AI, we strive to grow each employee to their full unique potential and we value and celebrate each person as they are. We are committed to diversity, equity, and inclusion. These values make us all more insightful and effective. Gaia AI does not discriminate on the basis of race, color, religion, sex, sexual orientation, gender identity or expression, national origin, age, disability, veteran status, marital status, or any other distinction unrelated to job performance. We especially encourage applications from minorities, women, people with disabilities, protected veterans, and all other qualified candidates.
We are looking for a purpose driven computer vision engineer to help build out our computer vision stack. This is an amazing opportunity for a computer vision engineer to join a well-positioned, ambitious technology startup whose products will disrupt and shape the forest carbon market.
Computer vision engineer responsibilities include developing the pipeline for processing imagery data, as well as training and deploying machine learning models to detect trees and classify tree species. To be successful in this role, you should have strong software development skills, experience with writing code for managing and processing imagery data, as well as have worked with machine learning models. If you also have experience in a graduate program working on machine learning models, or in building a computer vision pipeline in the context of a broader robotics tech stack, we’d like to meet you. Ultimately, you’ll unlock the true value behind camera data in verifying forest carbon credits within a tech stack that is uniquely set up to do exactly that.
- Make decisions on our camera hardware specs, and collaborate closely with the hardware team in choosing cameras and lenses.
- Write code that processes data from cameras and publishes it on ROS.
- Train and deploy machine learning models on data we collect from a forest.
- Refine the models with different state-of-the-art tricks, and track metrics to track what is improving the accuracy of the models.
- Read papers to stay up to date on state of the art approaches for machine learning applied to forestry data.
- Cooperate with our platform team and robotics team to integrate an effective computer vision pipeline into our product.
- Strong software development skills (will be demonstrated in your answers to the questions asked below as well as in the interviews)
- Experience working with machine learning and computer vision pipelines, ideally with forestry data or in the context of a robotics tech stack
- Background as a software developer
- Interpersonal skills for working well with a supportive team, and to be empathetic to customer needs
- Good time management
- Eagerness to learn / grow and yield results quickly
To apply, please answer the below questions and email your answers to Peter McHale at firstname.lastname@example.org, as well as your resume and why you are interested in the role. In order to consider candidates who are truly interested in what we are building at Gaia AI, applications without answers to these questions will not be considered.
If you use Chat GPT, you will not move on to an interview.
1) What data structures do you use the most and how do you choose between them? Illustrate with specific examples from the computer vision domain, avoiding vague generalities.
2) While C/C++ are famous for pointers and the bugs associated with them, understanding pointers is critical for effective programming in every language. Please provide a couple concrete examples of pointer-based bugs in scripting languages, ideally specific cases you have run into yourself.
3) We have just acquired a new sensor for a robot. Unfortunately, existing drivers are of very low quality so you need to write a new one. You only have access to the datasheet. How do you go about writing this driver? Please be both comprehensive and specific.
4) You've trained a deep neural network model and now need to evaluate its performance to see what makes it better as you make tweaks or add data. What are all the metrics you would use to evaluate its performance and why? Please explain these metrics to someone without a specialized background in CV.