Software Engineering MTS

Role details

To get the best candidate experience, please consider applying for a maximum of 3 roles within 12 months to ensure you are not duplicating efforts. Job Category Software Engineering Job Details About Salesforce 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. About the Team Accessibility is one of the hardest problems in software: it touches every component, every developer, and every user, and it's historically been slow, manual, and reactive. The Component Accessibility Team (CAT) is changing that. We improve accessibility by focusing on the developer experience: we build AI tools that shift accessibility as far left as possible, so developers don't have to be accessibility experts to write perfectly accessible code. We build the accessibility agent as a product : the knowledge, the fix tools, and the evaluation systems that let any Salesforce engineer request and receive accessible code suggestions in the flow of their work. Our tools ship into the surfaces developers already use: AI Expert Suite, DX MCP, Cursor, Claude Code, and Vibes. When a developer writes a component, we make it WCAG-compliant before the code makes it to production. This is a business problem, not only an ethical one. By improving accessibility at Salesforce we help win deals. It unlocks the features sales needs to close public-sector and enterprise contracts, and it protects Salesforce from a fast-growing wave of legal and regulatory risk. Accessibility serves the 1.3 billion people worldwide who live with disabilities, and it's increasingly a condition of doing business: ADA digital-accessibility lawsuits have increased every year since 2018, and VPAT conformance is now a standard requirement in enterprise and public-sector procurement. The US federal government can't buy software that fails Section 508, and the EU's European Accessibility Act extended the same bar to Europe in June 2025. Accessible-by-default code is how Salesforce keeps winning deals in the public sector, including with agencies that serve millions of veterans who have disabilities. The CAT accessibility product is already running at scale. Our Prizm PR-review presubmit is live on core and offcore and has reviewed 42,845 pull requests since late May. Engineers acted on 21% of its findings, and only 1% were ever reported incorrect. When this team gives an accessibility recommendation, it's almost always right, and it's landing in engineers' pull requests thousands of times a week. What the Team Is Responsible For Creating products that increase developer productivity — turning accessibility from a manual, after-the-fact audit into an automated, in-loop part of writing code. Guiding the company's web development efforts as the WCAG lead — ensuring compliance with accessibility guidelines and best practices across Salesforce's UI. Owning the full detect → diagnose → fix → verify loop: we don't stop at finding violations. We generate the fix, prove it holds on the rendered page, and drive it to a merged pull request. Ensuring AI solutions for accessibility are grounded in WCAG and internal standards — and heavily evaluated so that every suggestion the agent makes is trustworthy and high-performing. What You'll Build Well-evaluated AI fix tools — the Jest remediation tool, the functional-test (ftest) remediation tool, and the accessibility-guide tool that let developers request and receive accessible code suggestions based on the work in front of them. The agent's knowledge base — accessibility (a11y) best practices, remediation patterns, and Salesforce engineering standards that ground every suggestion in real WCAG success criteria. The evaluation system — the eval dataset and reviewer set (Aura + LWC) that prove, per WCAG criterion, that our AI performs before we ship it. If it isn't measured, it doesn't go out. The shared platform capabilities: component identification (mapping a rendered finding back to the exact component and source file it came from) and verification (confirming a semantic fix actually holds on the rendered accessibility tree). Detection for the dark corners of WCAG: first-ever automated coverage for categories like reflow (1.4.10) and focus management (2.4.3, 2.4.7) that no static check can reach. Packaging and adoption — getting the agent into the surfaces developers live in (AI Suite, DX MCP, Cursor, Claude Code, Vibes) and driving real usage, backed by the Iron metrics dashboard. What You'll Do Build and evaluate AI-powered accessibility tools Develop fix tools and agent skills that generate, validate, and repair accessible code across Aura and LWC. Design and expand the eval dataset and reviewer set so we can prove correctness per WCAG criterion — not just anecdotally, but with hard numbers. Treat evaluation as a first-class deliverable: no AI capability ships without a measured, defensible quality bar. Ground the work in real accessibility standards Encode WCAG success criteria and Salesforce internal standards into the agent's knowledge so suggestions are correct by construction. Expand per-criterion coverage and grow the reviewer set to close the gap between what tools can detect and what they can fix. Ship into developer workflows and drive adoption Package tools into the Claude surfaces developers already use, and remove friction so accessibility happens automatically, not as a chore. Remediate real Sa11y test failures (e.g., in Matrix) and feed what you learn b