Jobs at Icme
ICME builds PreFlight, a proactive AI security product that generates cryptographic proofs to verify that autonomous AI agents comply with defined rules before an action executes, rather than after the fact. It serves security and platform engineering teams shipping agentic systems that transact, approve, or spend money autonomously in agentic commerce, legal, finance, and enterprise risk settings.
ICME builds PreFlight, a proactive AI security product that generates cryptographic proofs to verify that autonomous AI agents comply with defined rules before an action executes, rather than after the fact. It serves security and platform engineering teams shipping agentic systems that transact, approve, or spend money autonomously in agentic commerce, legal, finance, and enterprise risk settings.
Maintainer signals as of 8/11/2026
About ICME
ICME is the company behind PreFlight, a guardrail system for autonomous AI agents that compiles policy rules into formal constraint problems and uses a solver to return a SAT (allowed) or UNSAT (blocked) verdict for each proposed agent action, in under one second. Every check produces a succinct, tamper-evident cryptographic proof that can be independently verified rather than trusted, replacing reactive, LLM-based safety judges with mathematical, proof-carrying enforcement. The product includes guardrails for prompt injection detection and PII leakage protection, and uses a proof-carrying-intent model (a variant of Incremental Verifiable Computation) to track and enforce a user's original intent across chains of orchestrator and sub-agent actions, blocking any agent whose behavior drifts from the committed intent. ICME targets CISOs, agent developers, and compliance teams at organizations deploying autonomous agents in agentic commerce, legal, financial, and insurance/liability workflows, offering an API-first product (PreFlight) accessible via documentation with no-credit-card trial access.
About ICME
ICME is the company behind PreFlight, a guardrail system for autonomous AI agents that compiles policy rules into formal constraint problems and uses a solver to return a SAT (allowed) or UNSAT (blocked) verdict for each proposed agent action, in under one second. Every check produces a succinct, tamper-evident cryptographic proof that can be independently verified rather than trusted, replacing reactive, LLM-based safety judges with mathematical, proof-carrying enforcement. The product includes guardrails for prompt injection detection and PII leakage protection, and uses a proof-carrying-intent model (a variant of Incremental Verifiable Computation) to track and enforce a user's original intent across chains of orchestrator and sub-agent actions, blocking any agent whose behavior drifts from the committed intent. ICME targets CISOs, agent developers, and compliance teams at organizations deploying autonomous agents in agentic commerce, legal, financial, and insurance/liability workflows, offering an API-first product (PreFlight) accessible via documentation with no-credit-card trial access.
No jobs published for this criteria in the past 3 months
Check back soon or explore all available positions