Staff Machine Learning Engineer, Salesforce

• 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
• 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
• 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

Desirable

• 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 dail
• 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
• 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
• 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
• 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

• End Date: June 30, 2026