AI ENGINEER · LINZ, AUSTRIA
VLADYSLAV DYKYI
I’m Vladyslav Dykyi. I build AI systems that connect models, data and tools — and keep working beyond the prototype.
01 / SELECTED WORK
A selection of the products and platforms I’ve helped build, from intelligent devices to multi-agent applications.
01 — EDGE AI & CLOUD
From the first model to the initial MVP.
As the first AI engineer on the LORAtech team, I developed the initial production-oriented MVP, connecting optimized edge models with cloud-hosted LLM services.
I worked across LoRA fine-tuning, MLflow experiments, quantized inference, backend services and Azure deployments — with guardrails, logging and tracing built in.
02 — RETRIEVAL & GENERATION
Natural-language mood requests, translated into Spotify recommendations. Vector retrieval, MMR diversification and constrained generation keep results grounded, with deterministic fallbacks when LLM calls fail.
03 — MULTI-AGENT ORCHESTRATION
A platform for creating and running custom AI agents. A LangGraph supervisor routes tasks, selects tools and coordinates agents, while SSE streams tokens, tool calls and hand-offs into the conversation.
02 / EXPERIENCE
Python automation, diagnostic services and APIs for production systems. I investigate failures, analyze manufacturing data with MongoDB, Elasticsearch and Kibana, and improve the reliability of diagnostic workflows across global operations.
First AI engineer on the team. Developed the initial MVP, trained and optimized models, built backend services, and deployed containerized workloads on Azure with Terraform and GitHub Actions.
Built a multi-agent platform with specialized assistants for biology, physics and mathematics. Designed RAG pipelines, Qdrant retrieval, PostgreSQL conversation memory and stateful LangGraph workflows with evaluation, tracing and recovery.
Delivered RAG systems, agents, APIs and automation. Built SOMA for a clothing business with 80k+ social followers, and WIZA, a travel assistant spanning 39 countries and 462 cities.
03 / TOOLKIT
RAG · LangChain · LangGraph · Tool calling · Structured outputs · Human-in-the-loop · Guardrails
MLflow · Evaluation pipelines · LoRA · ONNX Runtime · TensorFlow Lite · Quantization · Model testing
Python · SQL · TypeScript · FastAPI · Flask · NestJS · PostgreSQL / pgvector · Qdrant · MongoDB
Azure · Docker · Kubernetes · Terraform · GitHub Actions · CI/CD · Grafana · Kibana · Tracing
Databricks & GenAI. Hands-on, self-directed experimentation with notebooks, MLflow integration, vector search and model lifecycle workflows.
04 / A LITTLE CONTEXT
I’m an AI engineer based in Linz, Austria, focused on production GenAI. I enjoy the whole engineering problem: retrieval quality, agent behaviour, deployment, observability — and making all those pieces work together.
Get the full CV ↓EDUCATION
Johannes Kepler University Linz
Odesa National Economic University
LANGUAGES
CONTINUOUS LEARNING
Responsible AI & Generative AI · Coursera
Machine Learning with Python · Codecademy
Python Programming · Odesa Law Academy
NEXT / LET’S BUILD SOMETHING
Based in Linz, Austria.
Let’s talk about your next AI challenge.