11/06/2026 updated
Premium member
100 % availableAI Engineer | GenAI, RAG Systems & LLM Application Architecture
Tirana, Albania
Worldwide
Master of Science in Data ScienceAbout me
Lead AI Engineer & Founder with 4+ years building production GenAI systems, including RAG, semantic search, private chatbots, and AI agents. Experienced in LLM architecture, retrieval, evaluation, serving, and deployment. Led a 4-person team delivering B2B AI solutions for German enterprises.
Artificial IntelligenceMachine LearningTensorFlowMulti-Agent SystemsPrompt Engineering
GenAI & LLM Engineering
Production RAG pipelines, semantic search, private chatbots, AI agents. LangChain, Hugging Face Transformers, ReAct-style agents, tool/function calling, multi-agent systems, prompt engineering, LLM evaluation.
LLM Serving & MLOps
vLLM, TGI, llama.cpp, Docker, FastAPI, CI/CD, monitoring/logging, AWS. Full lifecycle from architecture to production deployment and ongoing support.
Retrieval & Data Systems
Vector databases (Milvus, pgvector), embeddings, semantic search, hybrid dense+sparse retrieval (OpenSearch), chunking strategies, RAG ingestion pipelines.
Unstructured Data Processing
Document parsing, OCR (Tesseract), PDF and scanned document processing, text extraction, chunking, metadata enrichment, document Q&A, multilingual document search, automated data-generation and evaluation pipelines.
ML / DL / NLP
PyTorch, TensorFlow, scikit-learn, spaCy, multilingual NLP, model training, tuning, and evaluation on real-world datasets.
Programming & Full-Stack
Python, SQL, JavaScript/React, Git. Building and deploying full-stack AI-powered applications.
Languages: English (fluent), German (fluent), Albanian (native)
Experience
Professional experience5 - 10 years
Freelancer since2024
Languages
GermanFluentEnglishFluent
Project history
Founded and leads a 4-person AI engineering team delivering production-grade GenAI solutions for German enterprise clients including subcontracted delivery for Innomatik AG. Architected and deployed GenAI applications including RAG, semantic search, and AI agents, owning the full lifecycle from requirements and solution architecture through production deployment and ongoing support.
Architected and maintained a production-grade RAG framework (BIT-RAG) with hybrid dense and sparse retrieval using Milvus and OpenSearch, LLM reranking, and automated data-generation and evaluation pipelines. Deployed open-source LLMs at scale using vLLM and TGI powering full-stack chatbot platforms with FastAPI and React, multi-tenant access control, multilingual support, and OCR-based document search.
Developed an evidence-grounded research assistant that retrieves relevant papers across multiple scientific databases and generates citation-backed answers using LLMs. Designed and built custom autonomous agents with reasoning, tool use, and self-correction capabilities based on ReAct-style architectures. Applied machine learning to process development, supporting data-driven decisions in pharmaceutical R&D workflows.