- 401(k)
- Health insurance
- Vision insurance
- Unlimited PTO
- Stock Option Plan
- Office food and beverages
Machine Learning Engineer - Palo Alto, CA, United States - Convectivecapital
Description
WindBorne Systems is supercharging weather models with a unique proprietary data source: constellations of next-generation smart weather balloons targeting the most critical atmospheric data. We then combine that unique data source with the world's most accurate AI weather models. Our long-term vision is to eliminate weather uncertainty, and in the process help humanity adapt to climate change, be that predicting hurricanes or speeding the adoption of renewables. The founding team of Stanford engineers was named Forbes under 30 and is backed by top investors including Khosla Ventures.
WindBorne is looking for a Machine Learning Engineer to work on our AI-based weather modeling. We're looking for generalists who love getting in the weeds and will be hiring multiple people over the course of the next year to work on this team. You will play a pivotal role in scaling up the world's most accurate medium-range weather models.
You will report directly to the CEO, John Dean, as he currently is running the AI team.
A balloon + flight paths from over 600 flights conducted by WindBorne
We're dedicated to both mitigating climate change and helping humanity adapt to its effects as fast as possible, and you can be a part of our mission to make this global societal impact. If you like the idea of working on cutting edge technology, and working with a fast moving team, WindBorne is the place for you.
Responsibilities
You will be working on all parts of our AI models: Architecting, Implementing, Training, and Evaluation. This role is ultimately fairly open-ended and responsibilities can somewhat match the specific interests of the candidate, so long as it aligns with the needs of the company and you are highly competent and willing to work hard. 95% of the machine learning stack is built by the CEO and another cofounder, and if you can do any aspect of it better than we can, you can own it. Our current high-level architecture works quite well, but it needs to be scaled up further and there are lots more fine-tuning and modification to training required. In addition, there is a lot of near-term work to be done in evaluating and running case studies of existing models. We also would like to publish parts of our work and get our metrics posted to public benchmarks. We want to continue to build new architectures and models to cover other areas of the weather prediction stack.
Skills and Qualifications
You must be fluent in Python and at least one compiled programming language, and generally be a strong programmer capable of holding a pure Software Engineering job. You must have a strong background in applied math generally, and specifically in information theory, statistics, and numerical optimization. You must have experience working with deep learning models, and you should have experience training large models but this is not an absolute requirement. You do not need any background in weather, but you must be highly interested and motivated to learn about numerical weather prediction and how weather works in general. Ideally, you are the kind of ML person who feels Fremdschämen when another ML person makes a rookie mistake while applying ML to some other field.