Staff Engineer
MCP Infrastructure + Security
About Ultra Security
Ultra Security is building the security foundation for the age of AI and MCP (Model Context Protocol). We help AI-native teams secure autonomous, agentic systems and earn customer trust without slowing product velocity.
Agents take action. Ultra secures them.
The Problem
AI teams today either move slow (manual approval processes, week-long waits for data access) or move fast and fail audits.
Ultra is building the AI layer that enforces security policies in real time -- agentic workflows that check compliance before data reaches a model, evaluation systems that catch policy violations, and infrastructure that auto-generates audit evidence.
This role sits at the boundary of applied AI and security infrastructure. There's no established playbook yet -- you'll help define the patterns.
The Role
We're hiring a Staff Engineer to sit "in the middle" of our Applied AI and Infrastructure lanes. We're looking for someone who is versatile and can reason deeply about distributed systems and runtime enforcement, while being equally comfortable partnering on AI-adjacent architecture (tool calling, evals, anomaly pipelines) when it intersects with security and compliance. Or knock out front-end application work when needed.
This is an execution-forward role with strong ownership. You will ship end-to-end capabilities across frontend/backend/services, data models, APIs, and product surfaces.
What You'll Do
Build product and infrastructure
- Help build Ultra's application and product end-to-end (frontend + backend), shipping polished user-facing workflows as well as the underlying systems
- Own new modules from 0 to 1 (and 1 to n): scope, design, implement, launch, and iterate
- Partner with the team to keep the product simple and fast while meeting enterprise security/compliance expectations
Ship quickly, iterate relentlessly
- Turn customer feedback into concrete improvements with tight loops: ship, measure, learn, repeat
- Operate with a bias to action and strong taste for pragmatic tradeoffs (you know when "good enough" is right, and when rigor is non-negotiable)
Build the core security platform
- Build and evolve the systems that make MCP safe in production: runtime enforcement, guardrails, auditability, and visibility
- Design for real-world constraints: latency, reliability, debuggability, and clear operator UX
Engineering craft + AI-native workflows
- Leverage modern AI tools day-to-day (Claude, Codex, Cursor, MCPs, etc.) to move faster, while maintaining correctness and rigor
- Contribute to evaluation/reliability thinking so new capabilities are testable and don't regress as the ecosystem evolves
What You Bring
Must-haves
- 5+ years building and shipping production software
- Staff/Principal/Founding-level trajectory (or equivalent scope/ownership)
- Strong backend/infra/systems background; proven distributed/data-intensive system design
- Security product experience (authz/access control, audit logging, compliance, runtime enforcement)
- AI/tooling adjacency: has built or integrated context/tool-bridge systems (MCP or similar) and can partner on eval/reliability thinking
- Startup or high-velocity environment experience (Seed to Series C)
Nice-to-haves
- Combination of backend development and infrastructure ownership within one role
- Experience as an early or founding engineer (first 10 employees)
- Work involving API-based or plugin-style integrations
Travel
Quarterly team offsites.