Software Engineer LMTS- AI/LLM

Job Category

Software Engineering

Job Details

About the role

Trust is Salesforce’s #1 value. Within the Salesforce Trust Platform (STP), the Trust Intelligence Platform (TIP) engineering organization builds the 360 security-intelligence layer protecting every Salesforce asset across our public-cloud substrates and private infrastructure. Inside TIP, the Trusted Supply Chain (TSC) team owns the services that secure the design, code, build, and run phases of the Salesforce SDLC - used by thousands of internal developers shipping the world’s largest SaaS platform.

As Lead MTS in TSC, you will own the architecture and engineering roadmap for our trusted-supply-chain platform: third-party-package security, vulnerability management, attribution, remediation, and runtime defense. You will be the authoritative voice on technical tradeoffs, the most active evangelist for next-generation security engineering, and the mentor who raises the technical bar for the engineers around you.

This is a builders role for a senior IC who wants to define what supply-chain security looks like in the agentic-AI era - not just instrument the SDLC, but redesign it.

What you will do

• Own the architecture vision and multi-year roadmap for Salesforce’s Trusted Supply Chain platform scanning, attribution, remediation, and runtime defense across public cloud and private infrastructure.
• Lead the AI-native rebuild of our supply-chain stack. Design and ship agentic systems that triage findings, author fixes, validate them in CI, and close the loop without human keystrokes for the well-understood cases - freeing engineers to focus on novel risk.
• Champion operational excellence. Feed incident learnings back into design. Build for fault isolation, graceful degradation, and observability from day one.
• Partner with PMs and EMs on build-vs-buy decisions and time-to-market vs. long-term sustainability tradeoffs.
• Coach Salesforce application teams on supply-chain and AppSec best practices; be the voice of TSC in the Salesforce Principal community.
• Track and adopt the most relevant trends in supply-chain security, agentic AppSec, ML-assisted vulnerability discovery, secure software development, and AI safety.
• Mentor engineers advocate for their contributions and career development; raise the technical bar across the team.

Required skills and experience

• M.S. or Ph.D. in Computer Science or equivalent practical experience.
• 10+ years building applications or systems software at scale.
• 5+ years in public-cloud environments with container-based technologies (Docker, Kubernetes, Istio, service mesh).
• Strong, active coding skills in at least two of Go, Java, Python, Rust, C#; with deep attention to code quality and secure-by-default patterns.
• Experience with stateful, service-oriented architectures and the design tradeoffs they impose (consistency, partitioning, recovery).
• Experience with relational and NoSQL stores and pub/sub messaging (Kafka, RabbitMQ, SQS).
• Demonstrated ability to define technology strategy and execute multi-year roadmaps through ambiguity.
• Track record of incremental architectural evolution, refactoring, and safely prototyping new features in production.
• Excellent written and verbal communication; proven ability to drive alignment with executives, peers, and engineers on contentious technical decisions.

AI / agentic engineering

This role is for someone who will build the next supply-chain stack, not just specify it.

We expect:

• Hands-on experience building production systems with LLMs - RAG pipelines, function/tool calling, structured output, evals, guardrails.
• Working knowledge of agentic patterns: multi-agent orchestration, planner/executor loops, tool-augmented reasoning, agent memory and context management, human-in-the-loop checkpoints.
• Practical fluency with the modern AI developer stack at least one of: Anthropic Claude / OpenAI / similar foundation-model APIs; agent frameworks (LangGraph, CrewAI, AutoGen, or equivalent); MCP (Model Context Protocol); evaluation harnesses (LangSmith, Braintrust, custom).
• Proven daily use of AI-assisted coding tools (Claude Code, Cursor, Copilot) including authoring custom agents, slash commands, hooks, and skills and a clear point of view on when to use them and when not to.
• A realistic posture on AI safety and reliability: prompt-injection defense, sandboxing untrusted tool calls, deterministic test harnesses, regression evals, cost/latency …