Slack is looking for a Staff Machine Learning Engineer with deep expertise in model training and finetuning to join our ML team. You'll design, train, and ship NLP models that power core product experiences - from summarization and search ranking to generative AI features used by millions daily. This role is hands-on: you'll work at a low level with training frameworks, optimize model architectures, build finetuning pipelines, and own the full lifecycle from experiment to production. At Slack, that impact can be huge: • We have over 10 million daily active users relying on our product. • At peak usage, a million messages a minute pass through Slack. • During the week, our users spend over a billion minutes a day active in our product. Machine learning engineers at Slack ship models that serve millions of users daily. This role owns that end-to-end: finetuning models for Slack's NLP tasks and putting them into production with the rigor and reliability our users expect. We're not looking for someone who hands off a checkpoint - we want someone who sees it through to serving traffic. Broader ML skills - data pipelines, experimentation, feature engineering - are valuable here too, but deep training and productionization expertise is the core of this role. This is a practical machine learning team, not a research team. Our goal is to deliver business value with machine learning and data in whatever form that takes. Sometimes that means bootstrapping something simple like a logistic regression and moving on. Other times that means developing sophisticated, finely tuned models and novel solutions to Slack's unique problem space. We are looking for engineers who are driven by driving impact for our business, building great products for our customers, and delivering robust, reliable services with machine learning. What you will be doing: • Design and execute finetuning strategies for large language models and other deep learning architectures tailored to Slack's NLP tasks (summarization, ranking, classification, generation). • Own the model training lifecycle end-to-end: data curation, training infrastructure, hyperparameter optimization, evaluation, deployment and monitoring. • Build and maintain scalable finetuning training pipelines on GPU infrastructure. • Brainstorm with Product Managers, Designers and Frontend Engineers to conceptualize and build new features for our large (and growing!) user base. • Produce high-quality results by leading or contributing heavily to large multi-functional projects that have a significant impact on the business. • Mentor other engineers and deeply review code. • Improve engineering standards, tooling, and processes. What you should have: • 5+ years of hands-on experience training and fine-tuning deep learning models in NLP (or a closely related domain like speech, IR, or multimodal). • 5+ years of experience with common deep learning frameworks like PyTorch, TensorFlow, JAX, etc • Track record of shipping fine-tuned models to production that serve real users at scale - not just research prototypes. • Experience with functional or imperative programming languages: PHP, Python, Ruby, Go, C, Scala or Java. • An analytical and data driven mindset, and know how to measure success with complicated ML/AI products. • Led technical architecture discussions and helped drive technical decisions within the team. • The ability to write understandable, testable code with an eye towards maintainability. • Strong communication skills and you are capable of explaining complex technical concepts to designers, support, and other specialists. Nice to have: • Expertise with recommendation systems or search. • Familiarity with model optimization for inference (quantization, pruning, speculative decoding, compilation via TorchScript/TensorRT/ONNX). • Experience with retrieval-augmented generation and hybrid retrieval/generation systems. • Broad experience across NLP, ML, and Generative AI capabilities. • Knowledge of using multiple data types in RAG solutions including structured, unstructured, and knowledge graphs. • Broad experience across NLP, ML, and Generative AI capabilities.