Job Post — RAG Engineer / Agentic Systems Engineer

Friendly, Clear, and Strict — “Who-first hiring” style (Claudio Fernández-Aráoz).

⸻

LeenAI — We Hire Builders, Not Buzzwords

Role: RAG Engineer / Agentic Systems Engineer

Location: Flexible — Remote or Dammam Office

Type: Full-time, Part-time, or Project-based

Start: Immediate (we move fast)

At LeenAI, we build practical AI agents for real Saudi businesses:

SmartQuote, OpsRAG, WhatsApp CX, Café Data Platform, Warehouse Vision, and more.

If you enjoy shipping real systems, debugging them on Friday nights, and taking ownership end-to-end — you’ll fit right in.

⸻

???? Mission of This Role (Simple & Direct)

Turn messy, unstructured business data into real, reliable AI agents that actually run inside companies — not on notebooks.

⸻

????️ What You Will Actually Do

(Not theory. Actual work.)

• Build RAG pipelines (chunking, embeddings, rerankers, evaluation).

• Create tool-based agents using OpenAI Agents SDK / LangGraph.

• Build connectors for PDF, DOCX, XLSX, Email, ERP, SharePoint, APIs.

• Implement guardrails (PII filters, relevance filters, HITL approval steps).

• Work on 6–8 week pilots with real customers and measurable KPIs.

• Write Python code that is clean, documented, and maintainable.

• Collaborate daily with LeenAI’s engineering + product team.

⸻

✔ Minimum Requirements (Non-Negotiable)

We are friendly, but we don’t compromise on these:

1) Proven RAG experience

You must have built at least one real RAG system:

• production or internal system

• not a tutorial

• not a copied template

2) Real agentic experience

You know how to build agents that use tools (API, DB, OCR, ERP actions).

Not just “chat completions.”

3) Strong Python

Async, FastAPI, debugging, writing tools, clean architecture.

4) Evidence of work

We do not accept vague CV claims — show us what you built.

⸻

❌ What We Do NOT Accept

• AI-generated answers

• Generic explanations

• “I have experience with all models”

• Fake projects

• Buzzwords without proof

If answers are not real → we simply won’t proceed.

⸻

???? Screening Questions

Friendly, but strict. Answer with honesty and details.

(Your answers reveal who you are — your character, judgment, and real experience.)

Q1 — Your Best RAG Project

Describe one RAG pipeline you personally built:

• What problem was it solving?

• What data sources did you ingest?

• How did you chunk the documents?

• What embeddings did you use?

• How did you evaluate quality?

• What was your exact role?

⸻

Q2 — Your Best Agentic Workflow

Describe one tool-using agent you built:

• What tools did it use?

• What decisions did the agent take?

• Show a concrete example of a full workflow.

• What parts did YOU build?

⸻

Q3 — Compliance & Data Handling

Tell us about a situation where you handled:

• PII

• Access control

• Logging

• Minimizing sensitive data

• On-prem or restricted environments

Explain your role clearly.

⸻

Q4 — A Failure You Learned From

Share a real story:

• What went wrong?

• How did you fix it?

• What changed in how you work today?

⸻

Q5 — Availability & Work Preference

• Remote or in-office?

• Full-time, part-time, or project-based?

• When can you start?

⸻

???? Our Culture (Friendly, Fast, Builder-Mindset)

• We move fast, but we don’t break trust.

• We care about craftsmanship, not job titles.

• We value honesty over perfection.

• We learn every week, not every quarter.

• We deeply appreciate people who show the work.

⸻

???? How to Apply

Send:

• CV

• GitHub

• 1 RAG project

• 1 Agentic project

• Answers to all screening questions

To:

راجع بيانات البريد الالكتروني على وظفنا دوت كوم