Thermo ML Resident

As a Thermo ML Resident at Extropic, your primary responsibility will be to collaborate with senior researchers in the development of new probabilistic models and their corresponding learning algorithms. This includes tasks such as scaling experimentation infrastructure, implementing and evaluating deep learning models, and contributing to cutting-edge research publications. You'll play an essential role in enhancing our understanding of thermodynamic paradigms in machine learning.

To qualify for the Thermo ML Resident position at Extropic, candidates should have experience in scientific Python and familiarity with deep learning frameworks such as JAX, PyTorch, TensorFlow, or Keras. A strong foundation in probability and linear algebra is crucial, along with practical experience in applied machine learning showcased through projects or published papers. Advanced knowledge in topics like energy-based models or graph neural networks will be advantageous but is not mandatory.

As a Thermo ML Resident at Extropic, the salary range is between $75,000 and $150,000 per year, reflecting the experience and qualifications of the applicant. In addition to a competitive salary, Extropic offers a dynamic work environment focused on innovation, collaboration, and professional growth. Equity compensation options may also be available, creating an appealing incentive for those looking to make long-term contributions to the company.

Absolutely! At Extropic, we encourage our Thermo ML Residents to contribute to academic discussions by publishing their research findings. Working alongside senior researchers, you will have opportunities to publish papers in reputable machine learning conferences such as NeurIPS, ICML, ICLR, and CVPR, allowing you to share your insights and innovations with the broader scientific community.

Preferred qualifications for the Thermo ML Resident position at Extropic include experience with energy-based models or diffusion models, as well as familiarity with graph neural networks or graph message-passing algorithms. Candidates with hands-on experience in deep learning infrastructure, such as using systems like Slurm, Ray, or Kubernetes for experimentation, and those who have a strong theoretical knowledge of information geometry and computational Bayesian methods will stand out in the selection process.

When answering, emphasize how probabilistic models allow for uncertainty modeling and decision-making under uncertainty. Discuss your understanding of how these models can be applied in various scenarios, specifically relating it to your experiences or projects you've worked on. Highlight your ability to derive new learning rules and the role of thermodynamics in shaping these models.

Reflect on your hands-on experience with these frameworks, discussing specific projects where you implemented models. Highlight the advantages you've found in using JAX or PyTorch, for instance, in terms of performance or ease of use. Be specific about the types of models you trained and how you optimized their performance.

Provide a narrative of a specific project, focusing on the challenges you faced, such as data issues or model convergence. Detail the solutions you implemented and what you learned from the experience, showcasing problem-solving skills relevant to the Thermo ML Resident role.

Discuss your methods for evaluating model performance, such as using metrics like accuracy or F1 scores. Explain how you visualize results using tools or libraries, like Matplotlib or TensorBoard. Share why visualization is crucial for understanding model behavior and making adjustments.

Demonstrate your knowledge by explaining the foundational concepts behind energy-based models, including their principles related to thermodynamics. Discuss diffusion models and how they can be applied in machine learning context. If possible, cite any projects or research where you've applied these models.

Highlight the importance of collaboration in driving innovation and fostering a learning environment. Provide examples from past experiences where you collaborated on projects and how these collaborations led to successful outcomes or breakthroughs in your work.

Discuss your strategies for keeping abreast of new developments, such as reading academic papers, following influential researchers on social media, or participating in relevant online communities. Explain how you incorporate this knowledge into your work as a Thermo ML Resident.

Explain the concept of Bayesian methods, especially with MCMC sampling and variational inference, highlighting their applications in uncertainty quantification and model optimization. Relate it back to how you'd use such approaches in the role of a Thermo ML Resident.

Discuss potential challenges like scaling models, managing computational resources, or staying ahead of fast-evolving ML technologies. Emphasize your preparedness to tackle these challenges with research, ingenuity, and collaboration with your team.

Describe your strategies for time management and organization, such as using project management tools, setting clear milestones, or prioritizing by impact. Stress the importance of maintaining quality and thoroughness even under tight deadlines.