Senior Software Engineer, Agent Engineering

Prophet Security•Menlo Park, CA

1d•Hybrid

About The Position

Prophet Security builds an AI-native security operations platform that accelerates detection, investigation, and response. We help security teams move faster without being constrained by alert volume or fragmented tools. As an early team, you'll work alongside veterans from Abnormal Security, Expel, Mandiant, and McAfee, and help shape what we build as we scale customer outcomes. Security operations is being fundamentally reshaped by AI agents. Prophet is building an autonomous AI agent for triaging alerts, investigating incidents, and recommending response actions; to do so, we need to resolve the challenge of ensuring our agent can effectively deliver outcomes in the messy reality of each customer's environment. You'll be the person who ensures Prophet's agent delivers these outcomes — debugging why a detection missed, figuring out why an integration is dropping events, building a workflow that connects the dots across a customer's security stack. Some days that means writing code. Other days it means getting on a call with a customer's SOC lead to understand why the agent's output isn't matching their expectations — and figuring out whether that's a data problem, a product gap, or both. This is not a support role. This is an engineering role where the customer's environment is your production system. Why Join Agent Engineering You're joining the founding Agent Engineering team. A few reasons this matters: Founder-level reps. You'll get reps in product discovery with real customers, shipping under real constraints, and developing the judgment to know when to build vs. push back – this is the same type of experience that helps you build the skillset to become a founder AI meets security, in production. You won't be reading about AI agents — you'll be making them work against real threats, in real customer environments No game of telephone. You work directly with SOC teams, see the problems firsthand, and ship solutions the same week. Your field work feeds directly into Prophet's product roadmap. It's a founding team. You'll help define how this function operates, not just execute within it.

Requirements

You write strong Python or Go and can scope, build, debug, and ship in production customer environments
You're energized by figuring out why something is broken. You can trace a problem end-to-end through unfamiliar systems, form hypotheses, and narrow things down
You're comfortable communicating with non-Engineering stakeholders, explaining technical tradeoffs, and building trust
You default to reusable over bespoke. When you build something for one customer, you're already thinking about whether it should become product
You've built Products or features where real users gave you direct feedback
You've built something 0→1, whether at an early-stage startup, as a founder, or standing up a new technology bet within a more established company
You’re always looking for the next challenge; the phrase “move fast, with conviction” resonates with you

Nice To Haves

Experience in security operations, detection engineering, or incident response
Familiarity with LLMs, prompt engineering, or agentic workflows
Experience with SIEM/SOAR tools, telemetry pipelines, or alert classification systems
Background building or maintaining API integrations or security connectors
Founder or early-stage (first 10 engineers) experience; thrives in ambiguity and ships fast

Responsibilities

Own end-to-end technical delivery for customers — from discovery and scoping through production deployment, ongoing iteration, and expansion
Debug Prophet's AI agent in production customer environments: figure out why detections miss, why triage logic misfires, why an integration setting is wrong — and fix it
Build fixes through prompt engineering, behavior design, integration configuration, or safer fallbacks — whatever it takes to make the agent work
Decide which platform primitives to assemble for a given deployment, and when bespoke work is genuinely required. When it is, time-box it and recycle it into reusable configurations or product features
Spot patterns across deployments and drive root-cause recommendations into Product and Engineering
Build quick integration workarounds to unblock deployments when needed, with a clear handoff to the Integrations team for anything that needs to be productized or maintained long-term
Partner with Customer Success, GTM, and Engineering to remove blockers, align on scope and tradeoffs, and keep execution moving