Machine Learning Scientist Intern

Job Title: Machine Learning Scientist Intern

Company: AppSofa (appsofa.com)

Location: Remote

Apply: Send your CV to hr@appsofa.com or apply directly through this platform.

Research Topics & Details: appsofa.com/research

About the Role

AppSofa is seeking highly analytical and motivated Machine Learning Scientist Interns to join our advanced research and development team. In this role, you will bridge the gap between theoretical machine learning and scalable, real-world execution. You will design core algorithms, conduct rigorous mathematical modeling, and push the boundaries of how complex systems learn and generalize.

We deeply value academic excellence and foundational contribution; as part of this internship, you will have the designation and support to co-author and publish your breakthroughs on arXiv and target peer-reviewed AI venues.

Core Responsibilities

Algorithmic Innovation: Design, mathematically formalize, and implement novel machine learning algorithms, optimization techniques, and foundational architectures.

Empirical Validation: Set up rigorous, reproducible experimental frameworks to evaluate model generalization, robustness, alignment, and scaling behaviors.

Research & Publication: Conduct deep-dive literature reviews, execute complex technical experiments, and draft high-quality scientific papers targeting arXiv and top-tier AI/ML conferences.

Internship Focus Areas

Candidates will have the opportunity to anchor their internship within a primary scientific track. Please indicate your preference for one of the following domains:

Advanced Alignment & Reasoning: Investigating novel reinforcement learning (RLHF/RLAIF), preference optimization, and test-time compute scaling techniques to enhance reasoning capabilities.

Relational Representation & Graph Transformers: Developing geometric deep learning frameworks and hybrid graph-transformer architectures to capture deep inductive biases in highly structured datasets.

Statistical Machine Learning & Optimization: Designing efficient optimization algorithms, exploring loss landscapes, and formalizing generalization bounds for specialized neural networks.

Multi-Modal Feature Synthesis: Advancing foundational representations across text, vision, and tabular data matrices to optimize multi-modal embedding alignment and zero-shot transfer learning.

Qualifications

Currently pursuing an MS or PhD in Computer Science, Applied Mathematics, Statistics, Electrical Engineering, or a highly quantitative field with an emphasis on machine learning theory and practice.

Advanced proficiency in Python and deep learning frameworks (primarily PyTorch).

Solid foundation in linear algebra, calculus, probability, and standard optimization techniques, with a strong grasp of modern deep learning theory.

Experience or documented interest in training pipelines, loss function customization, hyperparameter optimization, and statistical evaluation.

Track record of, or a highly demonstrated potential for, writing clear technical reports or contributing to academic manuscripts.

Apply: Send your CV to hr@appsofa.com or apply directly through this platform.

Research Topics & Details: appsofa.com/research