Director of Security Operations

Fluidstack•Austin, TX

1d•$250,000 - $350,000

About The Position

Fluidstack operates the compute infrastructure that powers frontier AI, including some of the most demanding training and inference workloads on the planet. We are building a Security Operations function from the ground up, and we want to build it right: AI-native, highly automated, and designed for the scale and threat model of a company that sits at the intersection of critical infrastructure and frontier AI development. The threat model here is not a narrow one. We operate corporate infrastructure and data center sites across multiple geographies, complex IT and OT/ICS environments, and cloud infrastructure, all serving customers whose work attracts sophisticated, persistent, and well-resourced adversaries. State-nexus actors, insider risk, supply chain compromise, physical intrusion, and infrastructure disruption are all real considerations. The SOC you build has to be credible against all of them, and the operating model has to hold up in a multi-stakeholder environment that includes upstream and downstream customers and partners with their own security requirements, audit rights, and contractual SLAs. This is not a role for someone who wants to manage a room full of analysts watching dashboards. This is a role for someone who wants to architect an entirely different model, one where AI handles L1 at scale, agentic workflows close the loop on routine response, a real threat intelligence function that drives detection, and where human analysts spend their time on work that requires genuine expertise and judgment. You'll be a builder across three dimensions simultaneously: the technical architecture, the operating model, and the team. If you've been frustrated watching the industry default to "hire more people" when the answer is "build better systems," this is the role you've been waiting for.

Requirements

You bring technical depth across the core disciplines
Proven experience designing or substantially rebuilding a SOC, not just running one someone else built
Deep hands-on background in detection engineering, SIEM/data lake architecture, and SOAR automation
Genuine experience with AI/ML applied to security operations, not familiarity with vendor marketing
Hands-on threat intelligence program development, including finished intel production and operationalization
Active threat hunting experience across heterogeneous environments
Exposure to OT/ICS environments or physical security telemetry at scale
Track record of reducing MTTD and MTTR through automation and architecture, not headcount
You know how to design an operating model, not just run one
Experience structuring coverage models, escalation logic, and stakeholder interfaces in environments where the org chart doesn't make things simple
Comfort navigating a multi-stakeholder environment with competing priorities and external accountability: customers, auditors, regulators
Experience operating under contractual security obligations with defined incident response SLAs
Ability to build processes that scale with automation rather than headcount, and to make that case credibly
You can lead a team and build a culture
Experience hiring, developing, and retaining security operations talent across a range of specializations
Ability to define team structure that matches the operating model: not the one that came before it
Track record of building culture in a function that operates under pressure

Nice To Haves

Experience with LLM integration into security tooling, including prompt engineering and evaluating AI output reliability under adversarial conditions
Data engineering fluency at the schema and query level
Experience designing SOC coverage for hyperscale or critical infrastructure environments
Threat intelligence program experience targeting sophisticated or nation-state-adjacent actors
Comfort in a compliance-adjacent environment (SOC 2, ISO 27001, FedRAMP-adjacent) without being compliance-driven

Responsibilities

SOC Architecture & Build: design and build FluidStack's security operations capability from scratch, including data architecture, detection logic, automation fabric, toolchain, and team model, using a modern stack
AI-Native Detection & Triage: define and implement a detection philosophy that assumes AI handles L1; build the pipelines, enrichment logic, and triage automation that resolves high-volume, low-ambiguity alert classes without human intervention
Agentic Response Workflows: design and deploy autonomous response workflows that contain, investigate, and remediate: not just notify; own and continuously push the boundary between machine-closed and human-required cases
LLM-Assisted Investigation: integrate LLM-based tooling into the analyst workflow for case summarization, log interpretation, and hypothesis generation; define how AI augments analyst cognition as a genuine force multiplier
Detection Engineering: own the detection content lifecycle end-to-end: MITRE ATT&CK coverage mapping, detection-as-code workflows, alert quality metrics, and continuous tuning across a heterogeneous environment
Threat Intelligence: build and operationalize a threat intelligence program that produces finished intelligence relevant to FluidStack's specific threat model and customer base, and connects directly to detection content and hunting hypotheses
Threat Hunting: design and run a proactive hunting capability operating independently of the alert queue, covering cloud, OT/ICS, physical telemetry, and endpoint across a threat landscape that includes sophisticated, targeted actors
Multi-Site Physical + OT/ICS Coverage: build detection coverage across data center sites, security-instrumented OT/ICS systems, physical access telemetry, and BMS environments that don't look like a standard enterprise
Operating Model Design: define the coverage model, escalation logic, stakeholder interfaces, SLA architecture, and feedback loops that make the SOC function as a system, not just a team
Team & Vendor Strategy: define the human layer of the SOC: size, structure, sourcing model, and skill profile; make the MSSP build-vs-buy call with data, not defaults
Customer & Regulatory Obligations: ensure the SOC can reliably and demonstrably meet contractual incident notification SLAs and compliance obligations across FluidStack's customer base