Enterprise AI Systems Engineer
for section headings. Use - for any list of duties/requirements/benefits. Use
for prose. Do NOT repeat job title as heading. Do NOT put Job ID etc. Remove boilerplate: EEO etc. Also remove "All applicants must have authorization to work for Juul Labs in the US." That's maybe legal; we can drop. Also remove "Juul Labs is proud to be an equal opportunity employer..." etc. Also remove "SALARY RANGES: Salary varies by role, level and location..." but we need to keep salary ranges if they are specific. The posting includes salary ranges per tier. So we should keep that. Also locations list.
We need to preserve all substantive information; do not invent or summarize away detail. Remove duplicated lines.
We need to format responsibilities as list items under a section. Similarly qualifications list. Benefits list. Education maybe just a paragraph.
We need to avoid sections with filler like "commensurate with experience" etc. Not present.
We need to output only HTML fragment.
Let's extract content.
First paragraph: THE COMPANY: description. That's about company mission. Could be placed under a heading? Not specified but we could include as a brief intro maybe under a heading "About the company"? Not required but we could include as a
before sections. The instruction: Use
for section headings (e.g. About the role, Responsibilities, Requirements, Qualifications, Skills, Benefits, Pay, Schedule). It says we can use those headings but not required to use exactly those; we can use headings that match content. Probably we should include a short intro about the company as a
before sections.
Then ROLE AND RESPONSIBILITIES: a long paragraph with many sentences each describing a duty. We should break into list items each duty. Let's parse sentences:
- You own these platforms end to end.
- What you build gets used across the whole company, and you decide how AI runs here rather than inheriting someone else's design.
- Administer our enterprise Claude, Gemini, OpenRouter, Cursor and NotebookLM tenants. User provisioning, groups and roles, workspace configuration, and tenant security settings.
- Build and maintain the content that makes those platforms useful: skills, prompts, projects, connectors, and the documentation behind them.
- Drive adoption across 250 users and growing.
- Track consumption and spend across all three platforms. Report on usage, find idle seats, and support the business-unit chargeback model.
- Keep the platforms running. Availability monitoring, feature rollouts, user issue triage, and escalation ownership when something breaks.
- Build and operate the MCP servers and gateways that connect these platforms to internal systems, with least-privilege access and full audit logging.
- Run the LLM gateway that centralizes routing, authentication, rate limiting, logging, and policy enforcement across model providers.
- Choose the right model for each workload, commercial or open-weight, based on cost, latency, capability, and how sensitive the data is.
- Build agentic workflows and automations against internal APIs and tools, with the evaluation and guardrails to keep them predictable.
- Architect and run AI workloads on AWS and GCP: compute, networking, IAM, secrets management, private connectivity, Bedrock, and Vertex AI.
- Own the code in GitHub and write most of it yourself, mostly Python and TypeScript. Branch protection, access control, secret scanning, and Actions pipelines.
- Instrument the stack for cost, performance, and security telemetry, and build the reporting that leadership and finance rely on.
- Enforce data handling, retention, and access policy alongside Cybersecurity and AI Governance, and write the runbooks and standards behind it.
- Perform related duties as assigned, within your scope of practice
We can turn each bullet into
- . Some sentences have multiple parts but okay.
Next PERSONAL AND PROFESSIONAL QUALIFICATIONS: list:
- 5+ years in software, cloud, or platform engineering, including recent hands-on work building and running LLM-based systems.
- Experience administering an enterprise SaaS or AI tenant at scale: access management, configuration, cost control, and adoption.
- Working knowledge of both commercial and open-weight LLMs, including prompt engineering, evaluation, retrieval, and agentic patterns.
- Hands-on with Model Context Protocol (MCP), MCP gateways, or LLM gateways. Comparable API integration and middleware experience counts.
- AWS and GCP architecture, including IAM, networking, secrets management, and managed AI services.
- GitHub and standard engineering discipline: version control, code review, testing, CI/CD, and infrastructure as code.
- Strong Python and/or JavaScript/TypeScript for building integrations, automations, and internal tooling.
- Security and data governance fundamentals applied to AI work: least privilege, data classification, DLP, and auditability.
- Able to explain technical work to executives, finance, and people who do not work in technology.
- Nice to have: Vertex AI or Amazon Bedrock, self-hosted open-weight model deployment, vector databases and RAG pipelines, Splunk or another SIEM, and cloud cost management.
We should separate required vs nice-to-have? Could keep as list, maybe note nice to have as a sublist or just include as item.
EDUCATION: "Preferred, bachelor degree in applicable field or relevant work experience". That's a sentence.
JUUL LABS PERKS & BENEFITS: list:
- A place to grow your career. We’ll help you set big goals - and exceed them
- People. Work with talented, committed and supportive teammates
- Equity and performance bonuses. Every employee is a stakeholder in our success
- Cell phone subsidy, commuter benefits and discounts on JUUL products
- Excellent medical, dental and vision, disability, and life insurance, plus family support, wellness, legal, and employee assistance program benefits
- 401(k) plan with company matching
- Plus biannual discretionary performance bonuses
We could keep as list.
SALARY RANGES: The text includes tiers and ranges. We need to present that info. Could be a
describing tiers and ranges, or a table? But we only allowed
- for lists,
for prose. Could use
with the details. Or we could make a
with each tier as - . That seems appropriate as list.
LOCATIONS: list of tiers with locations. Could also be a list.
We must not include boilerplate like EEO etc. Also remove "All applicants must have authorization to work for Juul Labs in the US." That's maybe legal; we can drop.
Now construct HTML.
Start with maybe a
for company description.
Then
Role and Responsibilities
...
Then Personal and Professional Qualifications
...
Then Education
Preferred, bachelor degree in applicable field or relevant work experience
Then Perks & Benefits
...
Then Salary Ranges
...
Then Locations
...
Make sure to not duplicate content.
Now extract exact text.
Company description: "THE COMPANY:Juul Labs's mission is to transition the world’s billion adult smokers away from combustible cigarettes, eliminate their use, and combat underage usage of our products. We have the opportunity to address one of the world’s most intractable challenges through a commitment to exceptional quality, research, design, and innovation. Backed by leading technology investors, we are committed to the same excellence when it comes to hiring great talent.We are a diverse team that is united by this common purpose and we are hiring the world’s best engineers, scientists, designers, product managers, operations experts, and customer service and business professionals. If the opportunity to build your career is compelling, read on for more details."
We need to keep that as prose. Remove "THE COMPANY:" label? Probably we can keep as part of paragraph but could strip label. The instruction: preserve ALL substantive information; do not invent or summarize away detail. The label "THE COMPANY:" is not substantive? It's just a heading. We could keep the sentences without the label. Safer to keep the sentences after colon. Let's just include the sentences.
Now responsibilities list items: need to ensure we capture each duty exactly as in text.
Let's copy each sentence:
1. You own these platforms end to end.
2. What you build gets used across the whole company, and you decide how AI runs here rather than inheriting someone else's design.
3. Administer our enterprise Claude, Gemini, OpenRouter, Cursor and NotebookLM tenants. User provisioning, groups and roles, workspace configuration, and tenant security settings.
4. Build and maintain the content that makes those platforms useful: skills, prompts, projects, connectors, and the documentation behind them.
5. Drive adoption across 250 users and growing.
6. Track consumption and spend across all three platforms. Report on usage, find idle seats, and support the business-unit chargeback model.
7. Keep the platforms running. Availability monitoring, feature rollouts, user issue triage, and escalation ownership when something breaks.
8. Build and operate the MCP servers and gateways that connect these platforms to internal systems, with least-privilege access and full audit logging.
9. Run the LLM gateway that centralizes routing, authentication, rate limiting, logging, and policy enforcement across model providers.
10. Choose the right model for each workload, commercial or open-weight, based on cost, latency, capability, and how sensitive the data is.
11. Build agentic workflows and automations against internal APIs and tools, with the evaluation and guardrails to keep them predictable.
12. Architect and run AI workloads on AWS and GCP: compute, networking, IAM, secrets management, private connectivity, Bedrock, and Vertex AI.
13. Own the code in GitHub and write most of it yourself, mostly Python and TypeScript. Branch protection, access control, secret scanning, and Actions pipelines.
14. Instrument the stack for cost, performance, and security telemetry, and build the reporting that leadership and finance rely on.
15. Enforce data handling, retention, and access policy alongside Cybersecurity and AI Governance, and write the runbooks and standards behind it.
16. Perform related duties as assigned, within your scope of practice
Now qualifications list:
1. 5+ years in software, cloud, or platform engineering, including recent hands-on work building and running LLM-based systems.
2. Experience administering an enterprise SaaS or AI tenant at scale: access management, configuration, cost control, and adoption.
3. Working knowledge of both commercial and open-weight LLMs, including prompt engineering, evaluation, retrieval, and agentic patterns.
4. Hands-on with Model Context Protocol (MCP), MCP gateways, or LLM gateways. Comparable API integration and middleware experience counts.
5. AWS and GCP architecture, including IAM, networking, secrets management, and managed AI services.
6. GitHub and standard engineering discipline: version control, code review, testing, CI/CD, and infrastructure as code.
7. Strong Python and/or JavaScript/TypeScript for building integrations, automations, and internal tooling.
8. Security
for prose. Do NOT repeat job title as heading. Do NOT put Job ID etc. Remove boilerplate: EEO etc. Also remove "All applicants must have authorization to work for Juul Labs in the US." That's maybe legal; we can drop. Also remove "Juul Labs is proud to be an equal opportunity employer..." etc. Also remove "SALARY RANGES: Salary varies by role, level and location..." but we need to keep salary ranges if they are specific. The posting includes salary ranges per tier. So we should keep that. Also locations list. We need to preserve all substantive information; do not invent or summarize away detail. Remove duplicated lines. We need to format responsibilities as list items under a section. Similarly qualifications list. Benefits list. Education maybe just a paragraph. We need to avoid sections with filler like "commensurate with experience" etc. Not present. We need to output only HTML fragment. Let's extract content. First paragraph: THE COMPANY: description. That's about company mission. Could be placed under a heading? Not specified but we could include as a brief intro maybe under a heading "About the company"? Not required but we could include as a
before sections. The instruction: Use
for section headings (e.g. About the role, Responsibilities, Requirements, Qualifications, Skills, Benefits, Pay, Schedule). It says we can use those headings but not required to use exactly those; we can use headings that match content. Probably we should include a short intro about the company as a
before sections. Then ROLE AND RESPONSIBILITIES: a long paragraph with many sentences each describing a duty. We should break into list items each duty. Let's parse sentences: - You own these platforms end to end. - What you build gets used across the whole company, and you decide how AI runs here rather than inheriting someone else's design. - Administer our enterprise Claude, Gemini, OpenRouter, Cursor and NotebookLM tenants. User provisioning, groups and roles, workspace configuration, and tenant security settings. - Build and maintain the content that makes those platforms useful: skills, prompts, projects, connectors, and the documentation behind them. - Drive adoption across 250 users and growing. - Track consumption and spend across all three platforms. Report on usage, find idle seats, and support the business-unit chargeback model. - Keep the platforms running. Availability monitoring, feature rollouts, user issue triage, and escalation ownership when something breaks. - Build and operate the MCP servers and gateways that connect these platforms to internal systems, with least-privilege access and full audit logging. - Run the LLM gateway that centralizes routing, authentication, rate limiting, logging, and policy enforcement across model providers. - Choose the right model for each workload, commercial or open-weight, based on cost, latency, capability, and how sensitive the data is. - Build agentic workflows and automations against internal APIs and tools, with the evaluation and guardrails to keep them predictable. - Architect and run AI workloads on AWS and GCP: compute, networking, IAM, secrets management, private connectivity, Bedrock, and Vertex AI. - Own the code in GitHub and write most of it yourself, mostly Python and TypeScript. Branch protection, access control, secret scanning, and Actions pipelines. - Instrument the stack for cost, performance, and security telemetry, and build the reporting that leadership and finance rely on. - Enforce data handling, retention, and access policy alongside Cybersecurity and AI Governance, and write the runbooks and standards behind it. - Perform related duties as assigned, within your scope of practice We can turn each bullet into
describing tiers and ranges, or a table? But we only allowed
- for lists,
for prose. Could use
with the details. Or we could make a
- with each tier as
- . That seems appropriate as list.
LOCATIONS: list of tiers with locations. Could also be a list.
We must not include boilerplate like EEO etc. Also remove "All applicants must have authorization to work for Juul Labs in the US." That's maybe legal; we can drop.
Now construct HTML.
Start with maybe a
for company description. Then
Role and Responsibilities
- ...
Personal and Professional Qualifications
- ...
Education
Preferred, bachelor degree in applicable field or relevant work experience
ThenPerks & Benefits
- ...
Salary Ranges
- ...
Locations
- ...
- . That seems appropriate as list.
LOCATIONS: list of tiers with locations. Could also be a list.
We must not include boilerplate like EEO etc. Also remove "All applicants must have authorization to work for Juul Labs in the US." That's maybe legal; we can drop.
Now construct HTML.
Start with maybe a