AI assistant (LLM/RAG) built into your SaaS by Irdi ZeneliAI assistant (LLM/RAG) built into your SaaS by Irdi Zeneli
AI assistant (LLM/RAG) built into your SaaSIrdi Zeneli
I build an AI assistant that works inside your product, not a demo: answers grounded in your own documents (RAG) with cited sources, tool calls into your API, a streaming chat UI, and evals so later changes don't silently break answers.
Recent work: Moza, an Albanian-language assistant prototype for law firms that drafts documents and exports them to Word for the lawyer to review; AlbSub, an open-source CLI that translates subtitles with OpenAI, Anthropic or Ollama and validates every line; and a voice-driven radiology reporting product used daily by a radiology team.
I work with OpenAI, Anthropic or open models, in TypeScript or Python. EU-based and GDPR-aware: your documents can stay on EU or self-hosted infrastructure.
I build an AI assistant that works inside your product, not a demo: answers grounded in your own documents (RAG) with cited sources, tool calls into your API, a streaming chat UI, and evals so later changes don't silently break answers.
Recent work: Moza, an Albanian-language assistant prototype for law firms that drafts documents and exports them to Word for the lawyer to review; AlbSub, an open-source CLI that translates subtitles with OpenAI, Anthropic or Ollama and validates every line; and a voice-driven radiology reporting product used daily by a radiology team.
I work with OpenAI, Anthropic or open models, in TypeScript or Python. EU-based and GDPR-aware: your documents can stay on EU or self-hosted infrastructure.