Avatar for Intuition Machines

Redefining the way companies apply ML to business

Reinforcement Learning Research Scientist

$50k – $100k • No equity
Apply now
Interested in working on hard problems at scale?

Help push forward the state of the art as part of our team of machine learning and distributed systems experts hailing from Apple, Amazon, Google, Stanford, MIT, etc.

IMI builds ML (Machine Learning) products and services used by some of the largest companies in the world. We have many problems that could yield to RL approaches, and millions of users (with the enormous datasets they produce) for some of our products.

As a RL specialist, we will lean on your expertise in defining problems, working with internal and external teams to attack these problems, and discovering novel approaches.

You should have a demonstrated academic or industrial research background in RL: please mention your relevant publications (and ideally point to your open source repos) to help us quickly get an overview your work.

You will likely work on both applied and theoretical problems at different times, based on your interests and the needs of the business. We have an international team from diverse academic backgrounds, and strongly believe in cross-pollination across projects, disciplines, and domains.

We also conduct original research in areas like unsupervised and active learning, motivated by the unique problems and datasets we have as one of the larger users of cloud resources for ML. Our work is often geared towards exploring applications of some of our more theoretical ideas, as we believe testing theories at scale is one of the best ways to push forward the field.

You will have the opportunity to publish at computer vision and ML conferences if your work yields results, and we are committed to open source code and (when possible) datasets: we regularly attend and publish our work at top conferences like ICLR, NeurIPS, CVPR, and more.

You have the option of working from our San Francisco, Berlin, or Helsinki centers, or remotely: we have a large distributed workforce working alongside the teams in our offices, and are comfortable with remote collaboration.

Research Interests

Experience and interest in one or more of the following areas is desirable:

Active Learning; applied Reinforcement learning
Time series analysis; temporal hierarchies
Few-shot learning; domain adaptation
Meta-learning; Neural architecture search; model compression
Theoretical analysis of Neural Network architectures
Distributed and Large scale training
Applied ML for Computer Vision tasks (Classification, Regression, Segmentation, Detection, Recognition) on Images and Video

In general, methods of learning from noisy or missing labels are all relevant areas, as we are in the fortunate position of having huge amounts of crowd-annotated data and even larger amounts of unlabeled data. We often work on methods to bootstrap systems using efficient large scale retrieval methods, and methods that can adapt to learn on multiple types of data.

In the context of data annotation, we are also investigating active and online learning, with the unique capacity to rapidly test and apply at scale ideas in this area.

Functional skills and experience:
We work primarily in Python for research, with pyTorch and Tensorflow being our preferred tools. You will have access to our in-house distributed training and inference infrastructure, which has been designed for ease of use but still benefits from basic knowledge of distributed systems. Some tasks may also involve helping to push forward our tools and libraries for large scale training, and transfer of results into our production systems.

We also strive for a high degree of programming competence within our research group, as we have found that good discipline in implementing ideas, along with good documentation, makes your task easier and collaboration more pleasant. If you are one of the many excellent researchers who have never written a unit test this will likely change!

As a researcher in a rapidly growing startup, your responsibilities will be fairly broad: we have a fairly flat structure and encourage everyone to wear multiple hats based on their talents.

Many of our researchers have PhDs or post-docs in relevant fields, often combined with industry experience. While not a strict requirement, having published in a relevant academic context is useful; but equally desirable is having a demonstrable codebase of quality projects. We appreciate that not everyone has prior work online, and completing a small test project with us is another way to demonstrate your skills.

Please send your CV, github, and a brief description of what most interests you, along with your dates of availability and preferred location.

IMI does not discriminate on the basis of race, creed, color, ethnicity, national origin, religion, sex, sexual orientation, gender expression, age, height, weight, physical or mental ability, veteran status, military obligations, or marital status.

More jobs at Intuition Machines

View all jobs

Machine Learning Research Fellow

Apply now

Machine Learning and Distributed Computing Internships

Apply now