AI is being run everywhere, by humans and agents. Harmonic Security governs all of it. The AI Governance and Control (AIGC) platform understands user / agent intent and data context in real time, giving security teams visibility and control across every tool and workflow so companies can move faster with AI while minimizing risk.

We operate in one of the fastest-moving spaces in tech: the emerging intersection of AI and cybersecurity. It's a market still being defined, and we're among the first to build natively for it - protecting sensitive data in real time, at the endpoint, with minimal friction for the teams using it. Enterprises get full visibility and control. Their people get to innovate without guardrails slowing them down.

Named to the Rising in Cyber 2026 list, Harmonic is gaining serious recognition and serious traction. We're led by deep cybersecurity expertise and backed by investors who know the space: N47, Ten Eleven Ventures, and In-Q-Tel.

As AI adoption inside the enterprise accelerates, the ability to safely observe, control, and enforce policy in real time isn't a nice-to-have. It's mission-critical infrastructure. That's what we build.

How We Work: AI-First by Design

Harmonic exists to help enterprises adopt AI safely and at scale. We hold ourselves to that same standard. Everyone here actively leverages AI tools to perform their best work, from deep research and writing to building robust processes and automating complex workflows. We expect every new hire to bring a genuine curiosity for AI and a commitment to using it to work smarter, faster, and with greater creativity.

For some, this involves tinkering and remaining open to emerging tools; for others, it means architecting entirely new systems with AI at the core. We will be transparent about expectations for every role and provide the tools and support needed for you to thrive.

About the Team

Our Product Delivery team is the engine that turns vision into impact. We ship early and often, getting valuable features into the hands of customers quickly and iterating from there. We work in the open by default, sharing progress and ideas, and we trust each other to own outcomes. We’re a small but mighty crew where every person plays a critical role and we’re committed to using AI to work smarter and faster.

About the Role

We're looking for an AI/ML Product Engineer to build the insight layer of Harmonic: the ML that makes sense of how people and agents are actually using AI, so our customers can see where it's creating value and where it's creating risk.

This is a hands-on, individual-contributor role with real product impact. You'll take problems from messy raw data all the way to a shipped, customer-facing insight, working directly with product, engineering, security and the customers themselves.

This is about turning AI usage into insight that customers act on. Classification and clustering are how we do that today, but you'll choose whatever approach fits the problem. It's Python-first engineering, not notebook analysis.

What You'll Do

Turn raw AI usage data into insights customers can see, trust and act on
Own insight features end to end, from framing the question with users and product partners through to production and iteration
Make sense of unstructured data at scale, with the quality of every output measured and understood
Build and maintain the data pipelines behind your models, and own their deployment and reliability in production
Build the evaluation frameworks, ground-truth sets and error analysis that prove an insight holds up before a customer sees it
Quantify uncertainty and communicate confidence clearly, so the product never presents a shaky signal as a certain one
Shape the product surface itself: what the model should show, to whom, and why
Keep pace with a fast-moving field, adopting new models and approaches as they prove themselves and retiring the ones that don't
Work across a polyglot stack: Python at the core, with Java (Spring Boot) and occasionally TypeScript when the task demands it, deployed on AWS

What You Bring

A track record of shipping production ML that customers or downstream product surfaces depend on, such as text or behaviour classifiers, or clustering and embedding pipelines
Solid applied ML foundations across supervised classification and unsupervised grouping, with a clear view of which model families you've used and why; this role isn't a fit for a background built mainly on forecasting or time-series prediction
Strong production Python for ML and data engineering – code in version control, tested and deployed, not just explored

AI/ML Product Engineer at Harmonic Security Inc