AI Engineer

AI ENGINEER

ABOUT ENT

Ent is the intent-aware workspace security platform for securing human and AI-driven work. Built to protect productivity, the new attack surface, Ent understands not just what users and agents do but why, and intervenes at the moment of risk before incidents occur. Where existing tools see events, Ent sees intent, so security teams can step in at the moment of risk instead of investigating days later. Founded by Lou Manousos and Brandon Dixon, co-founders of RiskIQ (acquired by Microsoft) and the team behind Microsoft Security Copilot, Ent is in production with Global 2000 customers across hospitality, financial services, and defense, and backed by Decibel, Sequoia, Crosspoint Capital, Craft Ventures, Shield Capital, Felicis, and In-Q-Tel. We’re now hiring the team that will define this category.

HOW WE WORK

Customer first. The product and the business are built around problems we’ve watched real security teams struggle with — not the other way around. Every roadmap conversation starts with what a CISO told us last week.

Humble. No drama. We hire people who share the mission and trust each other to deliver. Teamwork over showmanship. Accountability over politics. The work speaks louder than the person doing it.

Urgency. The window to build a durable security company in the AI era is open right now and it will not stay open. The shot clock has started. We move at the speed of the people we want to protect.

ABOUT THE ROLE

We are seeking a Senior AI Engineer to solve hard problems in running AI models where work actually happens: on the endpoint. As people and AI agents take on more of the work inside an enterprise, security has to understand behavior as it unfolds, not days later in an alert queue. At Ent, small, fast models on the endpoint watch the live stream of activity in context and make decisions in milliseconds. Larger reasoning models and agents in the cloud handle deeper investigation and response. You will take models and algorithms from our AI scientists and make them production-grade, with reproducible training, gated evaluation and efficient inference on the edge and in the cloud. Then you will own them in production, so the team can keep pushing into new research. This is a role for an engineer who wants to understand the models they ship, likes working without a map, and wants to help define how AI protects both humans and agents.

WHAT YOU’LL ACHIEVE (RESPONSIBILITIES)

• Take models and algorithms from our scientists into production, understanding them well enough to reimplement them correctly (for example Python to C++ or Java) and verify parity with the original.

• Own models after launch, including retraining, monitoring and evaluation, so researchers can move on to the next problem.

• Design evaluation and release gates that decide when a model is safe to ship, catch regressions from model or data changes, and explain shifts in metrics such as precision and recall.

• Build production monitoring for model quality, including drift detection, retrain triggers and ground-truth strategies for unlabeled data.

• Run inference within tight latency, memory, CPU and power budgets on the endpoint, and at scale in the cloud, using quantization, distillation, batching and caching.

• Build the pipelines that connect the two layers, so behavioral baselines and context learned in the cloud flow reliably back to endpoints.

• Build the infrastructure for training and experimentation (datasets, reproducible training, CI/CD for models) and keep it simple enough for a fast-moving team.

WHAT YOU’LL BRING (REQUIREMENTS)

MUST HAVE

• A track record of owning and designing whole systems end to end, shown by growing scope across roles or companies.

• Experience training, evaluating and deploying your own models in production. Integrating LLM APIs alone is not enough for this role.

• The ability to take someone else's algorithm, understand the math and logic behind it, and turn it int