Software Architect, CTO Lab

At JFrog, we’re reinventing DevSecOps to help the world’s greatest companies innovate. Our team of industry-leading software security experts is a true pioneer, constantly pushing the boundaries with original research and technological innovation. JFrog is a special place with a unique combination of brilliance, spirit, and just all-around great people. Thousands of customers, including the majority of the Fortune 100, trust JFrog to manage, accelerate, and secure their software delivery from code to production – a concept we call “liquid software”. Wouldn't it be amazing if you could join us on our journey?

JFrog manages the software supply chain for 80% of the Fortune 100 – packages, container images, ML models, agent skills, MCP servers, and AI-generated code. The rules are changing fast, and we are looking for strong engineers who can build the next generation of tools to manage, govern, and secure it all. JFrog's CTO Lab is a small, senior, multidisciplinary, innovative, and autonomous team building what comes next. We sit across the entire platform – Artifactory, Xray, Curation, AppTrust, JFrog ML, AI Catalog, Fly, Runtime, Distribution – and our job is to figure out how AI changes all of it. We run focused experiments, prototype fast, demo often, and grow what works into products alongside JFrog's product and engineering groups.

This role has two modes. In Build mode, you'll work across the full breadth of the platform – from artifact management and security to ML lifecycle and developer experience. In Scout mode, you'll be our antenna – evaluating new AI frameworks as they drop, scanning for emerging patterns in agentic AI, supply chain attacks, and developer tooling, and feeding evidence into JFrog's strategic decisions.

• AI-powered supply chain intelligence – Agentic systems that reason over artifacts, dependencies, and release signals to move past static block/allow rules toward decisions a senior engineer would make.
• Agent systems and governance – build and secure AI agents that operate against registries, pipelines, and deployment systems. Extend JFrog's controls to the AI artifact stack we ship today – MCP Registry, Agent Skills Registry, AI Catalog – and design the next generation of governance models.
• AI woven across the platform – versioning, security, and provenance for AI artifacts, and capabilities that make developers faster wherever speed, trust, or judgment can be amplified.
• Evaluation and measurement – benchmarks and pipelines that prove AI-powered approaches actually outperform traditional ones. Data beats opinions.
• Technology scouting and signal analysis – evaluate new AI frameworks and security innovations as they emerge. Concise assessments of what's real, what's hype, and what it means for JFrog.
• Engage with developers, architects, product leaders and customers. Present, analyze and lead architecture discussions.

• 7+ years building distributed, large-scale systems (or equivalent depth) – you've shipped production software, not just prototypes.
• Hands-on with AI/ML systems – you've built something real with LLMs, embeddings, RAG, or agent frameworks (LangGraph, CrewAI, AutoGen or similar). You understand prompt engineering, context windows, token economics, evaluation, agentic protocols (MCP, A2A and alike), and failure modes.
• Strong practical experience as a software engineer – you know how to write clean, modular code. 
• Independent Starter – you can independently produce demos and prototypes from scratch with a clear sense of where shipping begins.
• Judgment under ambiguity – you can take a hard, open question and come back with a working prototype and data. You'll kill your own project when the evidence says it doesn't work, and you'll be proud of what you learned.
• Technical leadership – you can analyze tradeoffs and guide teams, lead technical discussions and promote practical agreements with strong engineering practices. 

• Experience building AI agent systems – tool use, function calling, MCP, multi-step reasoning, sandboxing, and the security/governance of giving agents access to real infrastructure
• Hands-on with MLOps or ML model management – model registries, versioning, serving, monitoring, or security scanning
• Background in DevSecOps, supply chain security, or compliance – SBOMs, Sigstore, SLSA, OPA/Rego, DORA, FedRAMP, or package ecosystem internals (npm, PyPI, Maven, Go modules, Docker)
• Familiarity with JFrog's platform (Artifactory, Xray, Curation, AI Catalog, JFrog ML, JFrog CLI)
• Prior work in a research lab, innovation team, or early-stage startup where you built zero-to-one

The best person for this role isn't the one with the most impressive resume. It's the one who's already been building things like this on nights and weekends — and just needs the platform, the data, and the backing to do it full-time.

• Real problems at real scale - billions of artifacts, thousands of enterprise customers, and an AI transformation that's just getting started.
• Freedom to pick the right tool for the job - Claude, GPT, open-source models, fine-tunes, or classical ML. No vendor lock-in.
• A small team of senior peers who push back on your ideas and make them better.