Balun Energy
CTO of a Viennese Startup making world's best wind power forecasts with AI. ~3mio€² in funding, forecasting for 20%¹ of Austria's wind capacity.
About
I'm Yury: 10+ years shipping data projects; I bring a pragmatic, outcomes-first approach and love a good challenge.
I combine technical leadership (setting direction, unblocking teams, making tradeoffs) with implementation (models, data pipelines, deployment, monitoring).
I’m best when the problem is ambiguous, high-impact, and time-sensitive: turning “we should do AI” into a shipped product with measurable results.
How I help
Data Science & Machine Learning
- Forecasting, computer vision, NLP, and tabular models
- LLM-powered systems: RAG, vector, document Q&A, internal assistants, workflow agents
- Rapid prototypes
- Prductionize and integrate into your product (APIs, batch pipelines, dashboards)
Production & MLOps
- Data pipelines with traceability (batch + streaming)
- Reproducible ML workflows (CI/CD, versioning, testing)
- Scalable deployment (cloud/on-prem), cost/perf tuning
- Observability: quality, drift, latency, model performance in production
Tech Leadership
- Turn strategy into an execution plan: roadmap, milestones, success metrics
- Make hard calls: build vs buy, architecture, hiring, vendor selection
- De-risk delivery with crisp scoping, risk management, and technical reviews
Deciding what to build, or need someone to own delivery?
Book a 15-min fit call. I'll ask a few questions, then tell you whether I'm the right fit and what the fastest path looks like.
Projects and Success Stories
A selection of (non-NDA) projects
Anomaly Detection for European Steel Manufacturer
Transforming terabytes of sensor data into actionable insights for a leading European steel manufacturer. I developed machine learning models and real-time dashboards that identified critical furnace issues, saving millions of euros in scrap metal costs and establishing a permanent shop-floor monitoring system.
DailyLauncher
A customizable new‑tab experience that lets you design your own start page with drag‑and‑drop widgets.
Few-Shot Computer Vision for Industrial Wood Grading
Developed a log-face detection system for a leading European timber processor using just 50 training examples. My few-shot learning approach achieved 99%+ accuracy in production, enabling automated quality classification at conveyor speed.
Energy Forecasting for European Steel Manufacturer
Energy consumption forecasting for one of Europe's leading steel manufacturers with AI. In just days, I developed a proof-of-concept that achieved 97% accuracy for day-ahead predictions, enabling optimized electricity trading.
Creative Coding
A collection of generative art pieces exploring geometry, randomness and emergence using Processing.
Rail Cargo Austria Wagon Planning
Digitalising wagon allocation for a pan‑European rail freight leader. I worked with Rail Cargo Austria and research partners to turn a manual planning process into a data‑driven system that anticipates uncertainties in demand and generates optimal wagon plans within minutes.
MARLAnts
A multi‑agent reinforcement learning experiment where simulated ants learn to cooperatively build towers.
Superhuman Multi-Language OCR Engine
Building a custom document scanner outperforming Google Tesseract by 10x in accuracy and 100x in speed across 15 European languages simultaneously.
Restaurant Recommendation Engine
Building a production recommendation system from scratch using Singular Value Decomposition that delivered accurate personalized suggestions from sparse user rating data.
Lego Technic Mars Rover
A fully remote‑controlled 2.7 kg rover built from scratch with 11 motors, pneumatics and realistic functions.
Ricotta and Brunost
A kitchen experiment in making fresh ricotta cheese and turning the leftover whey into sweet brunost.
Latest Posts
Recent writing and thoughts
April 10, 2026
Context Engineering for Sovereign AI
Practical context engineering patterns for building reliable, EU-sovereign enterprise agents — from strict guardrails and structured retrieval to production tactics like retries, escalation, and lightweight evals.
February 25, 2026
Machine Learning on Big Data Workshop
Materials for the Machine Learning on Big Data workshop for Lumos Student DS Consulting
January 27, 2026
Architecting Reliable AI Agents for Production
What happens when your AI agent hallucinates a legal citation or a refund policy? Concrete ways to architect for reliability when deploying autonomous agents.
January 20, 2026
Are Knowledge Graphs in RAG better than regular vector RAG?
A simplified answer to when knowledge graphs add value to RAG systems versus when they just add unnecessary complexity.
Book a call
Send an Email
LinkedIn
GitHub