דרושים » הנדסה » Staff Research Engineer

LLM optimization is a young field with no settled playbook, so the optimizations you invent here are genuinely new - and they ship into production, not a paper. And they matter: real enterprises run enormous volumes of LLM traffic and feel every wasted token, so the work you do lands on live customer requests and solves a problem they actually have. It's about as close to the frontier, and as close to the product, as engineering gets.

03What you'll do
Invent and ship optimizations. Design new strategies across every layer of a request - prompt and context compression, tool pruning, file and dev-command trimming, semantic caching, cache-prefix stabilization - and take them all the way to the hot path. Each one has to cut tokens on real traffic without degrading output.
Prove every win. Build the evals and the token/quality instrumentation that separate a genuine saving from a plausible-sounding one. An optimization only ships when the numbers back it.
Live in the caching trade-offs. Provider prompt caches only pay off on byte-stable prefixes, and a naive rewrite can cost more than it saves. Knowing exactly when an optimization actually wins - deterministic transforms, per-session decision pins, semantic dedup - is a lot of the job.
Build and own the gateway that runs them. The real-time proxy that carries production LLM traffic - streaming, retries, backpressure - executing every optimization on the live request path.
Keep it fast and deploy anywhere. Defend a tight latency budget on the hot path, and package the whole thing so it installs cleanly into any customer's cloud - AWS, GCP, Azure, containers, Kubernetes - and upgrades without drama.
Own systems end to end. From first commit to code serving live customer traffic - plus the spend-visibility and governance product built on top - with a direct line to the founders.

The same systems work as our senior role, aimed at someone earlier in their career who gets real work done in a hard codebase with agentic tools. Systems programming has a reputation for being locked behind ten years of experience. We don't fully buy that. Someone with solid fundamentals, real curiosity about how machines work, and deep control of agentic coding tools can contribute here much sooner - as long as they have the judgment to check what the tools hand back. That judgment matters more here than almost anywhere: the agent runs as root and SYSTEM, and the codebase forbids unwrap, panic, and unchecked indexing, so you can't ship code you don't actually understand.

What you'll work on

Real pieces of the sensor across macOS and Windows: inventory, telemetry, the policy enforcement path, installers, and the integrations that hook into AI coding tools and browser extensions.

Moving through unfamiliar OS APIs quickly with agentic tools, then checking the result against the docs, the tests, and a real machine.

Small tools and tests that make the codebase easier for everyone to work in.

Tech you'll work with

The work is low-level systems development in Rust, close to the operating system across Windows, macOS, and Linux - memory, processes, files, and the OS internals that expose them. Alongside that, you'll lean hard on agentic coding tools; deep fluency with at least one is central to how we work, not a bonus. Expect to range across the stack - OS internals, AI tooling internals, CI, and a fair amount of reverse engineering - and to learn the parts you don't know yet.

The CTO Office is our core discovery engine, dedicated to navigating the rapid evolution in AI agent security and building the lighthouse. We are a high-agency team of ICs with integrated Engineering, Research, and Product mindsets. Our mission: incubate new ideas, define the approach to emerging attack classes, and build working POCs that prove technical reality. Guided by the CTO, Product, and GTM, we validate hard, open problems and actively support the Product organization by opening new development paths and seamlessly handing over advanced capabilities.

What Youll Do
Responsibilities:
Take a hard, open problem in AI agent security from this might be possible to a working POC that proves the company`s approach, grounded in technical reality.
Partner closely with Product, R&D, our Labs, and GTM. Dont just write about emerging problems. Think them through, design the solution, and POC it so customers see something real.
Own the end-to-end loop: scope the problem, build the prototype, evaluate it, document the approach, ship it.
Make the calls on architecture, tooling, and tradeoffs.
When a POC succeeds, own the handover to Engineering and Product. When a new direction needs its own team, lead the incubation.
Stay ahead of the rapid evolution in AI and proactively identify emerging trends.

A senior, hands-on engineer joining a small team that is already shipping a live platform. You own whole features end to end, from a rough problem to something running in production that a hospital relies on. You will not wait for a spec, and you help set the bar for how we build as we grow.
What you'll own
Whole feature verticals across frontend, backend, data, and cloud: the auditable AI record, the workflows around it, and the ingestion behind them.
The reliability of what you ship, including incidents on your own code.
Local architecture and tooling calls, made without escalating everything.
The engineering-quality bar a small team compounds on: review, testing, and the habits later hires inherit.

We are hiring Senior Forward Deployed Engineers - the Technical Builders, responsible for deciding that leads who design, build, and deploy the most complex agentic AI solutions directly inside enterprise customer environments. You'll lead technical delivery on your team, mentor the next wave of engineers, and set the engineering standards that define how agentic AI gets deployed at scale. Partnering closely with a Deployment Strategist (Strategic Builder - a customer-facing advisor who shapes AI strategy and drives adoption), you'll own outcomes end-to-end, from architecture decision to production. Your passion for well-crafted software that solves real problems will be supported by best-in-class AI tools and a team of talented builders around you.

The Senior Builder Experience

Technical Ownership: Partner with a Deployment Strategist to own customer outcomes end-to-end. You're the senior technical authority in the room, ensuring deliverables meet the highest engineering standards, whether it was generated by an AI tool by a colleague or written by you directly.

AI-Supported Engineering: Cursor, Claude, and Salesforce coding products like Vibes are embedded in your daily workflow. The conventions and patterns you develop tend to propagate through the team because other Builders look to your work as reference.

Roadmap Influence: Your field experience routes directly to Product and Engineering. The gaps you hit and the edge cases you encounter in real life customer environments become the features they ship. Senior Forward Deployed Engineers are the most credible voice in agentic AIproduct evolution.

Continuous Innovation: You're expected to stay at the forefront - piloting emerging AI tools, experimenting with new models and frameworks, and sharing what you learn with your team and customers.

we are using technology to transform transportation around the world. From changing a single persons daily commute to reducing humanitys collective environmental footprint - weve got huge goals.
As a Senior Full Stack Engineer, you will build the technology powering our companys Operation Center - the real-time heartbeat of our global service. This is a rare opportunity to own mission-critical systems that shape the daily lives of hundreds of thousands of riders. You will have full-stack ownership, working with a sophisticated stack and a team that values high-quality craft. We are looking for a high-agency engineer who leverages modern AI to outpace traditional workflows and is ready to grow into a formal leadership role.
About the Role:
End-to-End Ownership: Lead the development of our companys real-time operation center, managing complex routing, driver dispatching, and mass-scale user communication.
Full-Stack Architecture: Architect and ship features across our stack (React/TypeScript/Redux and Python microservices) with an emphasis on performance and scalability.
AI-Augmented Development: Integrate AI-driven tools (Copilot, Cursor, LLM-based testing/refactoring) into your daily workflow to increase velocity and maintain elite code quality.
Future Leadership: Mentor junior/mid-level engineers, drive technical roadmaps, and set the standard for engineering excellence within the team.
Strategic Collaboration: Work directly with Product and Design to translate ambiguous operational challenges into elegant, scalable technical solutions.

We're hiring for leverage, not years. If you can wield agentic coding tools better than most senior engineers - and you've shipped real, working software with them - we want to talk, regardless of how long you've been doing this or where you learned it. This role is designed for someone earlier in their career who, armed with the right tools and the judgment to use them, delivers like someone far more senior.

You'll work across the whole stack: a single feature might run from a Drizzle schema change, through a Hono endpoint, to a new React screen. The agentic toolchain is how you cover that ground - but the judgment to review, verify, and take responsibility for what the tools produce is what makes it work.

What you'll work on

Full, thin vertical slices - backend to frontend - shipped behind previews and a merge queue.

Extending and hardening our agentic-dev toolchain: skills, workflows, and MCP servers the whole team relies on.

Whatever's highest-leverage that week - you'll move across the codebase more than most.

Were looking for a Technical Program Manager who thrives in the middle of the action - someone who gets energized by complexity and turns it into clarity.
In this role, youll take complex, cross-functional engineering projects from messy beginnings to clean finishes, and drive every stage: scoping, planning, execution, delivery, and retrospective. You see AI not as a shortcut, but as a core part of how great program management gets done, and you actively design your workflows around it.
This is an AI-native environment - models ship fast, priorities shift, and what's possible changes week to week. You'll thrive here if ambiguity is fuel, not friction.
In this role, you will be responsible for:
Drive end-to-end ownership of complex, cross-functional engineering projects - from the first brainstorming session to post-launch retro
Design and supervise agentic workflows that automate status collection, risk surfacing, dependency tracking, and comms - so you operate at a higher level, not just faster
Build and manage AI-assisted program workflows that reduce coordination overhead and improve program visibility at scale
Spot bottlenecks before they become fires, and personally drive them to resolution
Map and manage cross-team dependencies so nothing blindsides anyone
Build and track key metrics that give real visibility into how engineering is performing
Apply sound judgment on escalations, stakeholder reads, and cross-team trust - the things AI surfaces but cant resolve
Identify opportunities for collaboration and innovation across teams - connecting dots others miss
Constantly communicate progress and status to stakeholders and management

We are constantly striving to make our systems reliable, scalable, and simple to operate so our services are available to travelers when they need them most. With our continued growth, we have exciting challenges ahead and we're looking for a Senior Site Reliability Engineer to join our team in Tel Aviv. This role blends classic SRE ownership with pragmatic AI SRE work: you will build and operate the platforms, automation, observability, and incident response practices that keep Navan reliable, while helping teams use AI solutions, AI providers, and their APIs safely and dependably.



This is a hands-on engineering role, not a research role. You will partner with product, platform, data, security, support, and incident response teams to make production systems and AI-powered experiences more resilient. You will use software engineering, infrastructure as code, SLOs, telemetry, provider observability, and automation as your main tools, and you will apply AI where it creates measurable reliability value rather than novelty.



This position is based out of our new Tel Aviv office.



What You'll Do:

Support AI-based application solutions where reliability matters. Partner with the development teams building AI-powered travel experiences to support the development and production operation of their solution.
Work with AI solutions, providers, and APIs. Partner with teams integrating AI capabilities and providers, with attention to API reliability, authentication, quotas, rate limits, latency and provider-specific operational constraints.
Troubleshoot AI tools and provider issues. Diagnose failures across AI-powered workflows, provider APIs, configuration, permission errors, degraded responses and related areas.
Operate reliable production platforms. implement and run cloud infrastructure,and help product teams move quickly without compromising reliability.
Improve observability. Build dashboards, alerts, traces, logs, and runbooks that make service health clear, actionable, and tied to SLOs and customer impact.
Apply AI to SRE workflows. Prototype and productionize AI-assisted systems that create effective and efficient operations
Automate operational toil. Create tools, workflows, and automation that remove repetitive manual work and make operational knowledge easier to use.