Browse categories
Explore
Fiverr Pro
English
$
USD
I will build ai chatbots, agents, and llm integrations
Erjon P
About this gig
Need an AI assistant that answers from your own documents, with real sources and no made-up facts?
For the past year, I've been a core engineer on a production AI legal assistant: lawyers chat with their case files and get streamed, cited answers drawn from tens of thousands of laws, court rulings, and knowledge bases across 5 jurisdictions.
What I can build for you:
- Chatbots and assistants trained on your PDFs, docs, website, or database
- RAG pipelines: smart chunking, embeddings, hybrid search and reranking
- GPT or Claude integrated into your app, with model routing to cut costs
- Data pipelines that keep your knowledge base in sync as documents change
- Integrations with databases, APIs, and tools like Slack and SharePoint
What you get:
- Answers grounded in your data, with citations and relevance filtering
- Lower API bills through prompt caching and cheaper models for simple tasks
- Production-ready streaming chat, error handling, and automated tests
- Full source code and clear documentation
Tech: Python, FastAPI, LangChain, OpenAI & Anthropic SDKs, Azure AI Search, Cosmos DB, Postgres, Docker, Azure
Not sure which package fits? Message me first.
Get to know Erjon P
Erjon P
Software Developer
- FromKosovo
- Member sinceOct 2026
Languages
English, Albanian
I am a software developer proficient in Python, Java, and TypeScript with a keen focus on AI development. As a Master’s student in Machine Learning and Data Science, I have hands-on experience automating workflows, building prototypes, and collaborating to solve real-world problems. I am passionate about creating innovative, user-centered solutions in dynamic environments.
FAQ
Which AI models do you work with?
OpenAI (GPT-5) and Anthropic Claude, including through Azure. I've built systems that run both behind one interface, so you can switch models or use a cheaper one for simple tasks without rewriting anything.
What do I need to provide?
Your documents or data (PDFs, Word files, website links, database access), a short description of what the bot should do, and an API key for OpenAI or Anthropic. I can help you set that up.
How do you stop the chatbot from making things up?
Answers come only from your content. I use hybrid search to find the right passages, a second AI check that filters out irrelevant results, and citations so you can verify every answer. If the answer isn't in your data, the bot says so.
My documents are long, messy or scanned. Is that a problem?
No. I've built ingestion pipelines for scanned PDFs, inconsistent formats, and JavaScript-heavy websites, and I split long documents into chunks that each make sense on their own. Very large files are processed in parts so nothing gets cut off.
What happens when my documents change?
I can set up automatic updates that re-process only the documents that changed. This keeps your bot current, avoids outdated answers and keeps embedding costs low.
What will the AI cost me each month?
It depends on usage. For most small and medium businesses, it's a few dollars to a few tens of dollars in API fees. I reduce costs with prompt caching and by sending simple tasks to cheaper models, and I'll give you an estimate before we start.
Can you add it to my website or app?
Yes. I can embed it as a chat widget on your site or deliver it as an API with streaming responses that your app can use. If you need it hosted, add the deployment extra.
Do you offer fine-tuning?
For most business use cases, RAG gives more accurate, up-to-date answers than fine-tuning, at a fraction of the cost. If your project really needs fine-tuning, message me and we'll discuss it.
