Role overview
- Train and fine-tune LLMs (e.g., LLaMA, Gemma, Qwen) for specialized tasks (such as function calling, charting, parameter parsing) and ensure their effective deployment across diverse environments, including on-device (MLX)
- Develop and apply model optimization and compression techniques (e.g., dynamic pruning, quantization) to enable high-performance on-device inference
- Conduct rigorous evaluation and benchmarking of LLMs to measure performance, accuracy, and efficiency across different setups
- Adapt LLMs for domain-specific tasks using techniques such as prompt engineering, adapter-based fine-tuning (PEFT), and multi-modal extensions
What we're looking for
- Strong background in classical ML and NLP
- Solid knowledge of deep learning, going beyond just LLMs (e.g., CNNs, RNNs, Transformers, Autoencoders)
- Proficiency in Python and deep learning frameworks (e.g., PyTorch)
- Deep experience with LLMs: training large and small models using various techniques (fine-tuning, post-fine-tuning, DPO, GRPO)
- Practical experience with LoRA / Q-LoRA or other adapter-based fine-tuning methods
- Hands-on experience in model optimization (pruning, quantization), evaluation, and deployment
- Ability to design and automate ML pipelines with a focus on scalability and maintainability
- Upper-intermediate English & fluent Ukrainian
- We are hiring talented humans. Meaning with all our variety of backgrounds and identities, including service members and veterans, women, members of the LGBTQIA+ community, individuals with disabilities, and other often underrepresented groups. MacPaw does not discriminate on the basis of race, color, religion, sex, sexual orientation, gender identity, national origin, veteran or disability status.
- Some benefits are under development, and new adjustments are possible.
Tags & focus areas
Used for matching and alerts on DevFound Fulltime Remote Ai Machine Learning Data Science