AI Research Scientist (LLM)

TL;DR

AI Research Scientist (LLM): Developing fundamental research for efficient on-device large language models in the macOS ecosystem with an accent on model optimization, architectural exploration, and localized machine learning. Focus on designing experiments, implementing algorithms from academic papers, collaborating with research labs, and transitioning validated prototypes into production.

Location: Ukraine or Europe; remote work or the Kyiv hub is available under a hybrid work model.

Company

A Ukrainian product company building consumer software, cybersecurity tools, and an ecosystem of products used by millions of people worldwide.

What you will do

• Prepare research proposals focused on LLM efficiency and optimization.
• Survey academic publications, formulate hypotheses, and run in-depth experiments.
• Collaborate with internal scientists and external academic laboratories on research projects and publications.
• Work with applied research specialists to validate prototypes and transition them into production.
• Drive an independent research niche from ideation through implementation.
• Contribute to scientific publications, paper clubs, and internal knowledge sharing.

Requirements

• Deep experience in NLP or a related machine learning domain.
• Strong theoretical and practical knowledge of LLM optimization, including quantization, KV-cache compression, speculative decoding, and distillation.
• Hands-on experience with parameter-efficient fine-tuning, including LoRA and adapters, as well as mixture-of-experts architectures.
• Ability to implement complex algorithms and architectures directly from academic papers.
• Advanced Python skills and experience with PyTorch, JAX, or TensorFlow.
• Strong foundations in linear algebra, probability theory, and mathematical statistics; upper-intermediate English or higher is required.

Nice to have

• Publications at international AI, ML, software engineering, or HCI conferences.
• Performance engineering experience, including profiling, GPU, Metal, or C++ optimization.
• Experience with localized LLMs and on-device frameworks such as MLX.

Culture & Benefits

• Remote work or access to a 24/7 Kyiv coworking hub with UPS and Starlink.
• Flexible working hours based on trust and cooperation.
• Medical insurance from the first month or an annual medical insurance allowance for employees abroad.
• Professional development opportunities, annual reviews, and internal communities.
• Personal time off, parental leave, and sabbaticals after five years.
• Veteran support, humanitarian initiatives, and financial assistance connected with the war in Ukraine.