Five Live AI-Agent Jobs I'd Actually Forward to a Builder This Week

Five Live AI-Agent Jobs I'd Actually Forward to a Builder This Week

Checked on May 6, 2026.

There are a lot of vague "AI jobs" lists floating around, but most of them mix together unrelated ML roles, stale reposts, and hand-wavy summaries. I wanted a tighter list: five roles with live application pages, direct apply links, and job scopes that are explicitly about agent systems, not just AI branding.

I also chose official ATS pages over recycled social posts. If a role is worth sending to someone serious about AI agents, the application page should be live and the scope should name concrete work like prompt evaluation, tool definitions, orchestration, context plumbing, observability, or production deployment.

Selection standard

I filtered for roles that met all four conditions:

1. The application page was live on May 6, 2026.
2. The role was remote or clearly online-friendly.
3. The work touched real agentic systems, not generic analytics or legacy ML only.
4. The apply link went directly to the company's ATS page.

The five-role shortlist

Role	Company	Work setup	Why it belongs on an AI-agent shortlist	Apply
Prompt Engineer	Netomi	Remote	Prompt design, tool descriptions for agentic frameworks, testing, benchmarking	https://jobs.lever.co/netomi/7fbf062a-4853-4336-a639-f2a607640d38
Forward Deployed Engineer	Aisera	Remote USA	Builds and deploys AI agents, integrates APIs/CRMs, uses prompt engineering, RAG, function calling	https://boards.greenhouse.io/embed/job_app?token=5630259004
Staff AI Engineer, Agent Orchestration	CookUnity	United States (Remote)	Multi-agent patterns, planning, memory, developer tooling, LLM-enabled SDLC automation	https://boards.greenhouse.io/embed/job_app?token=6645205003
Staff Fullstack Engineer - Grapevine (AI)	Gather	Remote	Builds context infrastructure, MCP and Slack interfaces, privacy/governance for AI agents	https://boards.greenhouse.io/embed/job_app?token=5633299004
AI Product Engineer, New Grad	Arize AI	Remote	Prompt and agent development tooling, eval infrastructure, troubleshooting agents	https://boards.greenhouse.io/embed/job_app?token=5396470004

1. Prompt Engineer at Netomi

Company: Netomi

Role: Prompt Engineer

Work setup: Remote

Direct application link: https://jobs.lever.co/netomi/7fbf062a-4853-4336-a639-f2a607640d38

What the posting is actually asking for

This is not a fluffy "AI enthusiast" posting. Netomi is hiring someone to:

• craft and refine client-specific prompts,
• define tool descriptions for agentic frameworks,
• automate prompt testing with scripts,
• build evaluation and benchmarking workflows,
• adapt prompts to customer business rules and model behavior.

The company positions itself as an agentic AI platform for enterprise customer experience, so the prompt work here is attached to real operational systems rather than toy demos.

Why it is relevant to AI agents

This role sits at the interface between model behavior and business logic. The important signal is not just prompt writing. It is prompt writing plus evaluation, tooling, and framework-level definitions. In practice, that is the difference between a chat gimmick and a deployable agent workflow.

Someone doing this job would be shaping:

• how agents interpret instructions,
• how tools are described to those agents,
• how outputs are tested before customer rollout,
• how prompt quality is benchmarked over time.

That makes it one of the clearest prompt-layer agent roles in the market.

Who this role fits best

A builder who is strong in NLP and Python, understands LLM behavior, and has already moved beyond one-off prompting into repeatable evaluation and benchmark design.

2. Forward Deployed Engineer at Aisera

Company: Aisera

Role: Forward Deployed Engineer

Work setup: Remote USA

Base pay range listed: $125,000-$175,000

Direct application link: https://boards.greenhouse.io/embed/job_app?token=5630259004

What the posting is actually asking for

Aisera is direct about the scope. The role includes:

• developing, configuring, and deploying AI agents,
• integrating those agents with APIs, databases, and CRMs,
• optimizing prompts and configurations,
• building with or around LangGraph or PydanticAI,
• working with RAG, function calling, and monitoring stacks such as Langfuse,
• translating customer requirements into production agent solutions.

This is classic forward-deployed work: take the model layer, the enterprise stack, and the business workflow, then make them behave like one system.

Why it is relevant to AI agents

A lot of agent roles stop at prototyping. This one does not. It is about deployment pressure: identity, integration, telemetry, debugging, and customer-facing implementation. That is where many agent systems fail in the real world.

The strongest signal in this posting is that it treats agent work as systems engineering, not content generation. OAuth, OIDC, JWT, APIs, RAG, and production refinement are all in scope. That is exactly the kind of role that turns agent theory into a shipped workflow.

Who this role fits best

Someone who can work across Python or JavaScript, enterprise integrations, and customer-facing implementation without falling apart when requirements get messy.

3. Staff AI Engineer, Agent Orchestration at CookUnity

Company: CookUnity

Role: Staff AI Engineer, Agent Orchestration

Work setup: United States (Remote)

Base pay range listed: $180,000-$200,000

Direct application link: https://boards.greenhouse.io/embed/job_app?token=6645205003

What the posting is actually asking for

CookUnity is hiring for a role that reaches deep into the orchestration layer. The posting describes work such as:

• building full-stack GenAI systems,
• operationalizing LLMs and agent frameworks across the software delivery lifecycle,
• developing multi-agent patterns for tool use, planning, and memory,
• embedding AI into PR summarization, code review, testing, and documentation retrieval,
• extending CI/CD with model-aware automation and observability.

This is one of the most technically explicit agent-orchestration roles in the current batch.

Why it is relevant to AI agents

The value here is breadth with real implementation detail. This is not just "build an assistant." It is about agentic automation across engineering operations: requirements, code generation, validation, post-deploy analytics, and telemetry.

In other words, the job is about designing a working agent loop around software delivery:

• context in,
• tool use,
• action planning,
• execution,
• measurement,
• feedback into the next run.

That is agent orchestration in the practical sense, not the conference-talk sense.

Who this role fits best

A senior engineer who has already shipped LLM features, understands evals and failure modes, and can connect backend systems, frontend tooling, and runtime monitoring into one coherent agent workflow.

4. Staff Fullstack Engineer - Grapevine (AI) at Gather

Company: Gather

Role: Staff Fullstack Engineer - Grapevine (AI)

Work setup: Remote

Base pay range listed: $188,700-$238,000 annually

Direct application link: https://boards.greenhouse.io/embed/job_app?token=5633299004

What the posting is actually asking for

Gather's Grapevine product is unusually clear about its mission: provide rich context so AI agents can navigate workplace tools such as Slack, Notion, and GitHub. The role includes:

• building ingestion for new external platforms,
• building privacy and governance systems so context does not leak between teams,
• shipping UI, MCP, and Slack interfaces,
• prototyping internal chatbots and workflow automation tools,
• working with AI systems, MCPs, and provider APIs.

Why it is relevant to AI agents

This role sits on the context layer, which is where many useful agents either become powerful or become dangerous.

A high-functioning work agent needs three things:

• access to the right knowledge,
• interfaces into the tools people already use,
• guardrails so it does not leak or misuse private context.

Grapevine is clearly trying to solve that triangle. The mention of MCP, governance, and company-specific knowledge systems makes this posting stand out from generic "AI app" roles.

Who this role fits best

A full-stack engineer who can move between ingestion pipelines, product surfaces, and governance constraints while still thinking clearly about what an enterprise-ready agent actually needs.

5. AI Product Engineer, New Grad at Arize AI

Company: Arize AI

Role: AI Product Engineer, New Grad

Work setup: Remote

Compensation listed: $100,000-$135,000 plus equity

Direct application link: https://boards.greenhouse.io/embed/job_app?token=5396470004

What the posting is actually asking for

Arize is best known for observability and evaluation, and the posting reflects that. The role includes work on:

• AI product innovation for teams building LLM applications,
• prompt engineering and agent development playgrounds,
• evaluation infrastructure at large scale,
• APIs and instrumentation,
• AI agents that help customers troubleshoot their own applications.

It is also explicit that new grads will be thrown into serious technical work early rather than sandboxed on low-stakes tasks.

Why it is relevant to AI agents

This is the evaluation-and-observability slot in the shortlist. That matters because agent systems do not become trustworthy just because they can call tools. They need measurement, debugging, failure analysis, and iteration loops.

Arize's role is attractive because it spans both sides of the stack:

• building tooling for prompt and agent development,
• and building infrastructure to tell whether those systems actually work.

For anyone early in career, that is a high-signal entry point into the agent engineering discipline.

Who this role fits best

A new graduate who wants to be close to production AI systems, is comfortable learning fast, and wants exposure to prompt tooling, agent behavior, and large-scale eval infrastructure rather than a narrow feature silo.

Why this list is stronger than a generic AI-jobs roundup

These five roles are not duplicates. Together they cover distinct parts of the agent stack:

• Prompt layer: Netomi
• Deployment and enterprise integration layer: Aisera
• Orchestration layer: CookUnity
• Context and interface layer: Gather
• Evaluation and observability layer: Arize AI

That spread makes the list more useful than a random collection of job titles with "AI" in them. A builder can look at this set and decide which part of the agent stack they actually want to work on.

Final take

If I were forwarding five openings to someone who says, "I don't want generic ML work, I want to build real agent systems," this is the batch I would send first.

Each role had a live application page on May 6, 2026, each one includes a direct application URL, and each one ties to a concrete agent capability: prompts, tools, orchestration, context, or evaluation. That is enough to make the list actionable, current, and worth revisiting even after the hiring cycle moves on.