Algorithms and Applications Team Intern

Job Summary Research internship focused on the design and discovery of next-generation algorithms for optimization, machine learning, and large-scale computational systems. The internship will contribute to the development of novel algorithmic frameworks that combine theoretical insights from statistical physics with automated algorithm discovery using machine learning and large language models. The Algorithms & Applications team at NTT Research explores new paradigms for algorithm design across the computing stack—from real-world optimization problems to machine learning architectures and emerging computing hardware. The group investigates fundamental questions about computational complexity, optimization dynamics, and the automated discovery of algorithms through AI-assisted methods. NTT Research is a global leader in advanced computing technologies and interdisciplinary research at the interface of physics, machine learning, and algorithm design. The team develops new algorithmic methods for challenging computational problems including combinatorial optimization, operational research, coding theory, neural architecture search, and hyperparameter optimization. A central theme of the team is the use of statistical physics principles to understand and reshape complex optimization landscapes. Current work includes techniques based on replica theory, cavity methods, and loop-expansion approaches to accelerate convergence to optimal solutions. In parallel, the team develops automated algorithm discovery frameworks leveraging machine learning and large language models. These systems treat algorithm generation as an evolutionary search process in program space. The team also explores hardware–software co-design, investigating how algorithm discovery techniques can adapt algorithms to emerging computing architectures. Interns will contribute to ongoing research projects and collaborate closely with scientists working at the intersection of algorithm theory, machine learning, and physics. Responsibilities Intern responsibilities may include one or more of the following research directions depending on background and interests. Optimization Theory and Statistical Physics Apply methods from statistical physics (replica theory, cavity methods, loop expansions) to analyze optimization landscapes. Investigate theoretical properties of average‑case algorithmic complexity and algorithmic dynamics. Combinatorial Optimization and Coding Develop new algorithms for combinatorial optimization problems arising in logistics, scheduling, and network design. Develop decoding algorithms and inference methods for next‑generation spatially coupled LDPC (SC‑LDPC) codes, leveraging ideas from replica coupling and statistical physics to improve decoding performance and scalability. Machine Learning and Algorithm Discovery Develop frameworks for LLM‑assisted algorithm discovery, combining evolutionary search, automated evaluation, and prompt engineering. Design scaffolding systems that connect LLMs to algorithm evaluation pipelines and optimization benchmarks. Qualifications and Skills Required Qualifications Exceptional candidates will be considered regardless of background. Currently pursuing a Master’s or Ph.D. in Computer Science, Applied Mathematics, Physics, Electrical Engineering, or a related field. Strong analytical and problem‑solving skills. Experience with programming (Python or similar languages). Interest in algorithm design, optimization, or machine learning. Strong written and verbal communication skills. Preferred Qualifications (depending on research focus) Optimization and Statistical Physics Knowledge of replica theory, cavity methods, spin glass theory, or average‑case complexity. Familiarity with graphical models, message passing algorithms, or loop‑expansion techniques. Operations Research and Optimization Knowledge of non‑convex optimization, combinatorial optimization, or heuristic methods. Experience with optimization modeling (e.g., logistics, scheduling, network optimization). Familiarity with coding theory or information theory. Machine Learning and Algorithm Discovery Experience with large language models and LLM APIs. Knowledge of prompt engineering, reasoning frameworks, or agentic AI systems. Familiarity with machine learning architectures and algorithmic search methods. Internship Details Duration: Up to 6 months Location: Sunnyvale, California (on‑site or hybrid depending on project) Level: M.S. or Ph.D. students Start date: Flexible #J-18808-Ljbffr