Data Scientist

Salesforce Salesforce · Enterprise · Bangalore, India, India

Salesforce is seeking a Senior Member of Technical Staff (SMTS) to lead the data science and software systems for predictive and agentic intelligence in Sales Cloud. The role involves designing, training, and deploying ML models for use cases like opportunity scoring and deal-risk detection, building and operationalizing agentic workflows with LLM-as-judge evaluations, and exploring new technologies like RAG and prompt tuning. The position also includes mentoring junior engineers and partnering with Product Management on roadmap development.

What you'd actually do

  1. Design, train, and deploy production ML models for core Sales Cloud use cases: opportunity scoring, deal-risk detection, and seller-facing explainable insights
  2. Build and operationalize agentic workflows on Agentforce, including LLM-as-judge evaluation frameworks and agentic evaluation metrics
  3. Partner closely with Product Management to help shape and drive the roadmap — bringing engineering perspective to prioritization and feasibility decisions
  4. Explore and adopt new technologies: improved forecasting methods, RAG architectures, prompt tuning strategies, and emerging AI tooling
  5. Mentor and coach junior engineers on the team

Skills

Required

  • Python
  • SQL
  • supervised and unsupervised machine learning
  • RAG applications
  • agentic workflows
  • prompt tuning and optimization
  • LLM evaluation
  • LLM-as-a-judge patterns
  • agentic evaluation metrics
  • communication and collaboration skills

Nice to have

  • Agentforce
  • Data Cloud / Data 360
  • Model Builder
  • Salesforce BDTs
  • Salesforce core codebase

What the JD emphasized

  • end-to-end data science and software systems
  • agentic intelligence
  • agentic workflows
  • LLM evaluation
  • agentic evaluation metrics

Other signals

  • end-to-end data science and software systems
  • predictive and agentic intelligence
  • building models that help sellers close deals faster
  • operationalize models
  • rigorous evaluations for agentic workflows
  • deliver explainable signals
  • agentic workflows on Agentforce
  • LLM-as-judge evaluation frameworks
  • agentic evaluation metrics
  • improved forecasting methods
  • RAG architectures
  • prompt tuning strategies
  • emerging AI tooling