תיאור המשרה
Above Security — Tel Aviv (5 days in office)
Above is building the AI-native insider risk platform for the modern enterprise — where humans and AI agents operate together inside security-critical systems.
We're backed by tier-1 global VCs, led by second-time founders, and already deployed with organizations operating at serious scale. AI is not a feature here — it's the system.
We're hiring a Senior AI Engineer to design, fine-tune, and operate AI agents and large-scale models in production. This role exists because off-the-shelf models aren't enough for the problems we're solving.
If you enjoy pushing models until they break — and then fixing them — keep reading.
What you'll do
Design and operate AI agents that reason, act, and collaborate with humans
Fine-tune and adapt large language models for:
Behavior analysis
Reasoning over long, messy timelines
High-precision enterprise workflows
Build agent orchestration systems (tool use, memory, planning, feedback loops)
Run large-scale inference and training pipelines in production
Work on model evaluation, drift detection, and continuous improvement
Optimize for latency, cost, and reliability at real enterprise scale
Partner closely with DevOps, security, and backend engineers — no research silos
Ship models that are auditable, explainable, and safe in sensitive environments
Hard requirements (non-negotiable)
5+ years in ML / AI / Applied Research roles
Hands-on experience fine-tuning large models (LLMs or multimodal)
Deep familiarity with agent architectures (tool use, memory, planning, reflection)
Real production experience
Heavy, daily usage of AI coding tools (Claude, Codex, Cursor, etc. — this is how we work)
Experience operating models at scale (high throughput, real traffic)
Comfortable working 5 days a week from our Tel Aviv office
Strong signals you're a fit
You've shipped agent systems that run unattended in production
You've fine-tuned models for precision, not just demos
You think about evaluation frameworks as much as training
You care about failure modes, hallucinations, and abuse cases
You prefer impact over papers
Nice to have (but not required)
Experience with RLHF / RLAIF / preference optimization
Background in security, fraud, or behavioral systems
Experience with multi-agent systems or long-running agents
Prior startup experience where scale arrived faster than expected
Why Above
You'll work on real agent systems, not wrappers around APIs
Problems are messy, adversarial, and high-stakes — perfect for AI
Tier-1 funding, serious customers, real scale
Second-time founders who ship
Office-first culture focused on velocity and ownership
Requirements
null
