# Software Engineering SMTS / LMTS - AI Core (REMOTE)

**Company**: Salesforce
**Location**: San Francisco, California
**Work arrangement**: remote
**Experience**: senior
**Job type**: full-time
**Category**: Engineering
**Industry**: Technology

**Apply**: https://salesforce.wd12.myworkdayjobs.com/en-US/External_Career_Site/job/California---San-Francisco/Software-Engineering-LMTS---AI-Core--REMOTE-_JR363384?utm_source=yubhub.co&utm_medium=jobs_feed&utm_campaign=apply
**Canonical**: https://yubhub.co/jobs/job_29084861-c26

## Description

As an Engineer on the AI Core team at Salesforce, you'll operate as a technical leader who turns ambiguous, high-stakes AI problems into durable, high-quality systems, while raising the technical bar for the broader team.

AI Core builds the retrieval systems and agent infrastructure that power how our AI agents find, trust, and use knowledge, and the developer tooling that lets engineers across the company ship production agents quickly and safely.

You'll lead significant initiatives end to end: faithful answering over a customer's own knowledge, retrieval that combines RAG and hybrid search, content lifecycle controls such as freshness, deduplication and source precedence, transcript processing, and the frameworks, registries, and evaluation tooling that make building a production agent a paved path.

**Responsibilities**

- Lead retrieval initiatives end to end: Design, build, and operate retrieval systems using RAG and hybrid search (lexical plus vector, reranking, filtering) so agents answer from the right content with the right amount of context.

- Build developer tooling for production agents: Drive technical design for the frameworks and platform pieces our engineers use to build agents, including multi-agent orchestration, shared primitives for prompts, tools, tracing, and evaluation.

- Own answer quality: Diagnose why answers are thin, wrong, or inconsistent by separating retrieval problems from generation problems (prompt, token budget, model choice).

- Design content and data models: Define how content from documents, knowledge bases, and call transcripts is ingested, segmented, labeled, and retrieved.

- Make AI changes safe to ship: Build evaluation, observability, and rollout practices (offline evals, online monitoring, regression detection) so we can change prompts, models, and retrieval with confidence.

- Set performance and scaling benchmarks across the platform we own: Define and track targets for retrieval and indexing, LLM provider usage, and agent orchestration.

- Own operational health: Ensure the reliability and performance of the systems your team owns.

- Use agentic development to raise leverage: Use coding agents by default for implementation, testing, refactoring, and documentation while holding a high validation bar.

- Partner across functions: Work with product, design, and engineering leadership to shape roadmap, tradeoffs, and requirements.

- Raise team effectiveness: Mentor engineers, give high-quality code review and technical guidance, and improve standards, tooling, and processes.

**Requirements**

- Proven experience shipping production AI or backend systems in a high ownership environment, with a recent focus on LLM-powered applications.

- Expertise building retrieval systems using RAG and hybrid search, including chunking and indexing strategy, embeddings, vector and lexical search, reranking, and measuring retrieval quality.

- Expertise building developer tooling that lets developers build production agents.

- Experience owning platform initiatives from design through adoption.

- Demonstrated technical leadership: you lead significant projects, design solutions for non-trivial problems by weighing tradeoffs, and influence architecture and execution beyond your immediate workstream.

- Strong track record delivering complex projects with multiple stakeholders, dependencies, or phases.

- Strong backend engineering fundamentals and proficiency in languages such as Ruby on Rails, Python, JavaScript, or Go, with SQL and RDBMS experience (PostgreSQL preferred).

- Hands-on experience with the AI application lifecycle: experiment setup, dataset curation, offline evaluation, error analysis, deployment, and online evaluation.

- Comfortable operating in ambiguity and driving clarity: you can frame problems, propose solutions, and align others.

- Strong communicator across functions, able to explain technical ideas to both technical and non-technical audiences and to translate complex concepts into business impact.

- Ownership mindset and bias for action.

- Curious and continuously learning, with an agent-forward approach to how you build software.

## Skills

### Required
- RAG
- hybrid search
- LLM
- Python
- Ruby on Rails
- JavaScript
- Go
- PostgreSQL
- SQL

### Nice to have
- pgvector
- query performance at scale
- content ingestion
- evaluation suites
- voice or real-time conversational agents
- enterprise SaaS
- Salesforce/CRM integrations

---

Source: [Apply at salesforce.wd12.myworkdayjobs.com](https://salesforce.wd12.myworkdayjobs.com/en-US/External_Career_Site/job/California---San-Francisco/Software-Engineering-LMTS---AI-Core--REMOTE-_JR363384?utm_source=yubhub.co&utm_medium=jobs_feed&utm_campaign=apply)