Principal Software Engineer / Data Scientist, AI

Salesforce•San Francisco, CA

5d•$97,300 - $344,700

About The Position

Salesforce is the #1 AI CRM, where humans with agents drive customer success together. Here, ambition meets action. Tech meets trust. And innovation isn’t a buzzword — it’s a way of life. The world of work as we know it is changing and we're looking for Trailblazers who are passionate about bettering business and the world through AI, driving innovation, and keeping Salesforce's core values at the heart of it all. Ready to level-up your career at the company leading workforce transformation in the agentic era? You’re in the right place! Agentforce is the future of AI, and you are the future of Salesforce. The SDB AI team builds the foundation that makes agentic engineering on Salesforce Database safe, correct, and reusable across the org. Salesforce Database is the multi-tenant production database behind Salesforce's cloud platform no general-purpose model knows. We encode that database-specific knowledge, prove that agents acting on it behave correctly, and take agentic work all the way to production.

Requirements

Hands-on experience building and shipping full-stack, full-lifecycle agentic AI systems in production – LLM orchestration, tool and MCP integration, retrieval and knowledge grounding, and evaluation.
Strong programming in Python and Java in a Unix/Linux environment; solid grasp of distributed systems.
Judgment about what makes agents reliable at scale: grounding, evaluation, and guardrails against the failure modes – hallucination, context drift, wrong remediation – that cause incidents.
Track record of shipping and operating production software at scale, with a bias for hands-on delivery over specification.
Strong written and verbal communication; able to drive AI discussions across engineering teams and influence without authority.
You should have 10+ years of professional experience.

Nice To Haves

Familiarity with the modern data and agent stack a plus: tracing and experiment tracking (e.g. MLflow), workflow orchestration (e.g. Airflow), distributed query engines (e.g. Trino), knowledge graphs, and model-agnostic gateways across frontier models.
Experience building automation for root-cause analysis and self-remediation of production systems is a strong plus.
Background and experience in Data Science highly desirable.
Experience with large-scale distributed databases or multi-tenant SaaS architectures is a strong plus.

Responsibilities

Build the agentic memory the agents reason over – the knowledge and context graphs that hold the database schema, runbooks, query templates and escalation-policy – and keep it accurate as the database evolves.
Build closed-loop evaluation as infrastructure: local tracing, replay and golden-trace regression, PR-gating, and scoring against domain-grounded rubrics – proving not only that output is correct but that the execution path is.
Advance the production agent from workflow-shaped to agent-shaped: a dynamic, causal planner and the specialized tools database workflows need beyond off-the-shelf MCP.
Take agentic work all the way to production – through the platform, identity, and integration layers where the real bottleneck lives – and operate it at machine speed across hundreds of incidents a week with high-accuracy root-cause analysis, capacity planning, and early detection of anomalous events – with high-accuracy outcomes.
Build the guardrails and telemetry that keep agents grounded: context-drift detection, activation instrumentation, and failure-mode analysis.
Partner with database engineering teams to turn one team's pattern into every team's starting point – paved paths, deduplication, and quality gating that convert org-scale sprawl into one trustworthy library.