Common questions.

How do I start a project with you?

Email me at artenisalija@gmail.com with a brief description of what you're trying to automate or build. I'll reply within 24 hours to set up a discovery call — usually 30–45 minutes. No commitment required.

Do you work remotely?

Yes, entirely. All my clients are remote. I'm based in Tirana, Albania (UTC+2) and regularly work with clients across Europe, the UK, and the US. Async communication works well for most projects.

How do you price your work?

Most projects are fixed-price after scoping. This means you know the cost upfront and I'm motivated to deliver efficiently. For ongoing work or projects with evolving scope, I offer a monthly retainer. I don't do hourly billing for new clients.

How long does a typical project take?

A focused automation (one n8n workflow, one scraper, one integration) typically takes 1–2 weeks. Larger builds — LLM pipelines, multi-agent systems, full data infrastructure — range from 4–8 weeks. I always give a timeline estimate after scoping.

Will I be able to maintain the system after delivery?

Yes — maintainability is a design goal, not an afterthought. I write documentation, record a walkthrough video for handover, and structure code so your team can modify it. For n8n workflows especially, the visual interface makes non-technical maintenance accessible.

What's the difference between n8n and a Python script?

n8n is a visual workflow tool — great for multi-step orchestration where the shape of the process changes often or needs to be understood by non-developers. Python scripts are better for heavy data processing, custom logic, or when you need the full power of the language. I often use both in the same project.

What's the difference between automation and AI?

Automation follows fixed rules: if X, do Y. AI makes decisions: given X, determine the best Y. Most real-world systems I build combine both — automation handles the orchestration and data movement, AI handles the decisions (classification, generation, scoring) within that flow.

Can you integrate with my existing tools?

Almost certainly. If it has an API or a webhook, I can integrate with it. n8n has 400+ native integrations. Beyond those, I build custom nodes or direct HTTP calls. The tools I encounter most: HubSpot, Salesforce, Notion, Airtable, Slack, Google Workspace, Stripe, Shopify, and most SaaS platforms.

How do you handle sensitive or private data?

I default to self-hosted infrastructure when data sensitivity is a concern — your data never leaves your server. For LLM calls specifically, I offer local model deployment (Ollama + Llama 3) as an alternative to cloud APIs for clients where data cannot leave the premises.

Do you work with non-technical clients?

Yes. A significant portion of my clients are non-technical — founders, operations managers, sales teams. I scope and communicate in business language, not code. The deliverables are documented and where possible managed through tools (like n8n) that are accessible without engineering background.

Which AI models do you work with?

Primarily OpenAI's GPT-4o and GPT-4o-mini for production work. Claude (Anthropic) for tasks where instruction following and long context matter. Local Llama 3 / Mistral via Ollama for privacy-sensitive deployments. I choose based on the task, not preference.

How reliable are AI agents in production?

Reliable if built correctly — which means confidence thresholds, human escalation paths, structured output schemas, and rollback logic for destructive actions. I don't deploy agents without undo paths. The engineering around the model matters more than the model itself.

What is RAG and do I need it?

RAG (Retrieval-Augmented Generation) connects a language model to your documents or knowledge base, so it answers based on your data rather than general training. You need it if you want an AI system that accurately references your specific content — product docs, support history, internal policies.