Applied ML Scientist / Senior Applied ML Scientist (Natural Language Processing)
OpenTable
- Vancouver, BC
- $132,000-188,000 per year
- Permanent
- Full-time
- Conduct research and development of Natural Language Processing (NLP) models, and collaborate with engineers to bring these models into production. The responsibilities include, but are not limited to:
- Utilizing prompt engineering and parameter-efficient fine-tuning methods to enhance the performance of Large Language Models (LLMs) for a specific task.
- Constructing data pipelines essential for the training and evaluation of NLP models.
- Implementing comprehensive evaluation suites for measuring and assessing the performance and safety of models on a specific task.
- Develop applications that utilize large language models (LLMs), enhancing their capabilities through RAG and through incorporation of different Tools.
- Contribute to the internal ML packages, and help the team to build tools for training, evaluating, debugging, and interpreting NLP models for retrieval, reranking and generation.
- Stay current with NLP research, know when to apply it to your work, and how to communicate it to your partners.
- Experience applying transformer model architectures to NLP challenges is crucial, we expect you to have significant experience experimenting with encoder and especially decoder architectures.
- Experience applying parameter-efficient fine-tuning for LLM as well as familiarity with different training techniques SFT, DPO, PPO.
- Experience with optimizing LLMs for inference
- Strong understanding of data structures, algorithms, and OO design
- Strong knowledge of Python and hands-on experience with NLP-related /scientific computing packages (HuggingFace ecosystem (Transformers, TRL, PEFT), DeepSpeed, PyTorch, NumPy, SciPy).
- Passion for continuous learning and self-development
- Strong communication skills and the ability to work with others in a closely collaborative team.
- Contributions to an open-source Machine Learning (ML) package, showing your skills in researching, implementing, and evaluating academic papers.
- Publications in the field of NLP.
- Hands-on experience in deploying Large Language Models (LLMs) to real-world products, particularly in environments sensitive to latency where the model processes many queries simultaneously.
- Proficiency in Large Language Model (LLM) inference deployment, with knowledge in relevant technologies and packages, such as ONNX, FasterTransformer/TensorRT-LLM, llama-cpp, Triton Inference Server and VLLM.
- Participation in Kaggle competitions focused on NLP, demonstrating your in-depth understanding of problems and data, as well as your ability to experiment with a diverse set of NLP techniques / models to find an effective solution.
- Experience in building Retrieval-Augmented Generation (RAG) or other applications leveraging Large Language Models (LLMs).
- A graduate degree or equivalent in Machine Learning, Mathematics, Statistics, Physics, Economics, or a related technical field is preferred.
- Paid Time Off - 20 days a year
- Birthday/celebration PTO - 1 day
- Flexible sick time off
- Paid volunteer time
- Annual company week off
- Parental Leave Benefits
- Dental & Vision Insurance
- Life & Disability Insurance
- Group RRSP and DPSP
- Major Medical Insurance (dependent care options)