Job Description

The Visa Services

is the UAE's leading visa agency of 2025 (World Travel Awards). We help clients travel without borders by delivering visas for Schengen, the USA, Japan, China, and many other destinations--combining deep expertise, speed, and premium service.

At the same time, we're building an in-house

IT team

and developing a product that automates visa and migration services. Our technology is first implemented internally to improve efficiency and transparency, and then scaled into standalone

SaaS products

for the international market.

Due to fast growth and the expansion of our

AI direction

, we are looking for an

AI/ML Engineer

who will help us build reliable analytics, turn data into management decisions, and become a core contributor to a product transforming the migration services industry.

What You Will Do

Build Agents:

Create and deploy autonomous AI agents (LangChain/LangGraph) that can use tools, maintain context, and handle API failures gracefully.

Build Production RAG:

Design search pipelines (Hybrid Search, Re-ranking), select vector databases (Pinecone/Weaviate), and fight hallucinations to ensure the agent relies on facts, not fiction.

Tune Models:

Fine-tune transformers (Llama, Mistral) for sales and dialogue specifics where prompt engineering is not enough.

Measure Quality:

Implement evaluation systems (DeepEval, RAGAS), because "it seems to answer okay" is not a metric. Set up Human-in-the-loop processes.

Turn Chaos into Structure:

Work with real data (calls, chats), clean it, and prepare it for training.

Must-have Requirements

AI Agents & LLMs:

Hands-on experience building and shipping AI agents using LangChain (or similar). You know how to make LLMs use tools (tool calling), manage memory, and handle failures.

Transformers:

Practical experience working with transformer models (training, fine-tuning). You understand how to handle context window constraints in real systems.

Evaluation:

Experience implementing LLM/RAG evaluation using DeepEval, RAGAS, and retrieval metrics (e.g. Recall@K). You know how to build Human-in-the-loop (HITL) workflows.

RAG:

Production experience with embedding models and vector databases (FAISS / Pinecone / Weaviate). You can design chunking & overlap strategies, hybrid retrieval, and grounding to reduce hallucinations.

Good-to-have (Bonus)

MLOps:

Experience deploying ML/LLM systems to production. You are familiar with model & prompt versioning, monitoring (latency, cost, errors), and CI/CD for ML pipelines.

Data Engineering:

Experience building and maintaining data pipelines for text data: ingestion, cleaning, deduplication, PII handling, and dataset versioning.

Software Development:

Strong software engineering background. You write production-grade Python code, design APIs, write tests, and conduct code reviews.

Why This Is Interesting

Real Tasks:

We work with a live stream of leads (6k+ per month). Your agents will impact real business revenue.

Freedom and Impact:

You will define the architecture of the product's AI core.

Focus on Results:

Minimum bureaucracy, maximum common sense and shipping features.

What We Offer

Full official employment in the UAE (employment visa + medical insurance). Competitive salary. A friendly work environment and culture that values collaboration and professional growth.
Job Type: Full-time