AI & LLM Engineering

From bare metal
to running inference.

RAG pipelines, self-hosted models, vector search, and the infrastructure to run them — designed by someone building with the same stack.

$ ollama pull gemma4:e4b
pulling manifest ████████████████ 100%
$ docker compose up -d qdrant postgres
Container qdrant started
Container postgres started
$ curl localhost:11434/api/tags
{"models": ["gemma4:e4b", "nomic-embed-text"]}
$
View projects About me

What I do

End-to-end LLM delivery

From provisioning the server to shipping the RAG API — the full stack, handled with precision.

01

Infrastructure & Setup

VPS provisioning, Docker orchestration, Coolify deployments, DNS management, SSL, firewall hardening. The server work most teams skip until it breaks.

02

Self-Hosted AI

Ollama on bare metal, model selection by hardware constraints, Qdrant vector stores, embedding pipelines. Private inference that doesn't bill per token.

03

RAG Development

Production-grade retrieval-augmented generation: hybrid search (dense + sparse), multi-provider LLM connectors, streaming, audit trails, and evals.


Built in production

Recent projects

Systems built to run, not to demo. Each ships with Docker, CI, tests, and real auth.

finance_rag ↗ github

Financial Document RAG API

Upload financial PDFs and query them in natural language. Hybrid search merges Qdrant vector results with PostgreSQL full-text via Reciprocal Rank Fusion. Multi-provider LLM connector supports Ollama, OpenAI, Groq, DeepSeek, Gemini and more. Audit trail with faithfulness and relevance scoring via Celery.

FastAPI Qdrant PostgreSQL Celery Authentik Ollama React
books_rag ↗ github

Virtual Library RAG

Ask any question across an ingested book library. APA-cited answers with exact page references. Handles PDFs without real page numbers via font-size heading detection. Incremental ingestion — only reprocesses changed books. Grafana + Loki observability stack included.

Flask Qdrant SQLAlchemy Alembic Grafana Loki React
document-copilot ↗ github

Document Copilot

Intelligent document assistant for enterprise workflows. Conversational interface over structured and unstructured document sets. Built with multi-provider LLM support, streaming responses, and vector-powered retrieval.

Python LLM Vector Search Streaming Docker

Technology

The stack

Chosen for production reliability, not trend-chasing.

Inference
Ollama
Vector DB
Qdrant
Relational
PostgreSQL
API
FastAPI
API
Flask
Background
Celery
Containers
Docker
Deployments
Coolify
Frontend
React
Language
Python
Language
TypeScript
Observability
Grafana
Auth
Authentik
Models
Gemma 4
Embeddings
nomic-embed
Proxy
Traefik

Serious about the craft.

I build production LLM systems — from the server they run on to the API that serves them. No shortcuts, no toy demos.

Currently pursuing a Masters in Business Information Systems, with a research focus on applied LLM systems in enterprise contexts. The academic side informs how I think about system design; the production side keeps it grounded.

The niche is deliberate: AI infrastructure is where most LLM projects fail. Models are the easy part. Embeddings, retrieval quality, hybrid search tuning, latency at scale — that's where the work is.

3+
Production RAG systems built and deployed
10+
LLM providers integrated (Ollama, OpenAI, Groq, DeepSeek, Gemini, Mistral…)
MBiS
Masters in Business Information Systems — in progress
E2E
Infrastructure to inference — the full stack, not just the model layer

Let's build something.

RAG system, self-hosted AI setup, or LLM infrastructure consultation.