AI Security Engineer
Keeps prompts, data, tools, and agents inside the lines, and provably so.
Secures the new attack surface AI creates: prompt injection, data exfiltration through context, tool misuse, model access governance, and the audit trail your enterprise buyers will demand.
PROVABLY INSIDE THE LINES
WHAT THEY OWN
Concrete deliverables, not job-description poetry.
01
Prompt injection defense
Layered controls for direct and indirect injection, tested adversarially.
02
Data boundary enforcement
PII and secrets kept out of prompts, logs, embeddings, and fine-tunes.
03
Tool-use security
Least-privilege permissions and sandboxing for anything an agent can touch.
04
AI red teaming
Structured attacks on your own system before someone else runs them.
05
Access governance
Who can call which model with which data, enforced, logged, reviewable.
06
Compliance evidence
SOC 2 / ISO / EU AI Act-ready documentation of your AI controls.
PRICING
Pick the level, keep the senior oversight.
Junior
$2,600 /month
or $16/hr on Time & Material
AI-native from day one
Mid-Level
MOST HIRED$3,800 /month
or $24/hr on Time & Material
Independent feature ownership
Senior
$4,800 /month
or $30/hr on Time & Material
Architecture & judgment
Dedicated engineers are billed monthly; Time & Material is billed hourly on tracked actuals. The free trial week applies to every dedicated hire.
YOU NEED THIS ROLE IF
An enterprise deal just asked how you secure your AI features
Agents can call tools and nobody wrote down which ones
Customer data flows into prompts and nobody has mapped where it goes
BY END OF WEEK ONE
Threat-modeled your highest-risk AI path
Ran a first injection and exfiltration test suite
Locked down the scariest tool permission
Drafted the controls doc your buyers keep asking for
OUTCOMES YOU CAN MEASURE
Security answers that close enterprise deals
Agents with least-privilege, audited tool access
Injection attempts that fail and get logged
AI features your CISO stops worrying about
PAIRS WELL WITH