Burak Kağan Yılmazer
Data & AI
Engineer
About
ML & Creative Technologist with a Data & AI Engineering backbone.
Author of Pixels2GenAI , an open-source curriculum that walks learners from raw pixels to GANs, VAEs, and Diffusion Models, now growing into a broader platform for creative AI education.
Outside the terminal, I paint, build real-time visuals in TouchDesigner, and exhibit installations that blend evolutionary algorithms and generative models with live audiences.
Skills
- Python
- SQL
- Rust
- dbt
- Airflow
- FastAPI
- REST APIs
- ETL/ELT
- Data Modeling
- Anthropic Claude API
- Claude Agent SDK
- MCP (FastMCP)
- LangGraph
- Langfuse
- RAG (pgvector)
- Ollama
- Claude Code
- PyTorch
- GANs
- VAEs
- Diffusion Models
- TouchDesigner
- ONNX Runtime
- MediaPipe
- TensorFlow
- scikit-learn
- NumPy
- Pandas
- OpenCV
- YOLO
- AWS
- Docker
- CI/CD
- Git
- PostgreSQL
- Snowflake
- DuckDB
- Power BI
- Metabase
- Tableau
Experience
- -
Data Analyst → Data Engineer @ Bookingkit
Berlin, Germany - On-site · Part-timeSummary:
- Joined as Data Analyst building Metabase dashboards and SQL/DBT models, then promoted to Data Engineer to design and optimize ELT pipelines with DBT, Airflow, and Python.
Responsibilities:
- Design and optimize ELT pipelines using DBT, Airflow, and SQL
- Automate data workflows and monitoring with Airflow and Python
- Develop and maintain data models in DBT
- Design interactive Metabase dashboards and analyze business KPIs
- Implement data quality monitoring and alerting systems
- SQL
- Python
- DBT
- Airflow
- Metabase
- -
Senior Advanced Analytics Analyst @ Medtronic
Prague, Czechia - On-siteSummary:
- Develop, inspect, and model data to improve productivity. Conduct statistical analysis for predictive forecasting. Manage end-to-end data processing.
Responsibilities:
- Develop, inspect, and model data to improve productivity
- Conduct statistical analysis for predictive forecasting
- Manage end-to-end data processing
- Maintain analytical systems and verify data accuracy
- SQL
- SAP ERP
- SAP S/4HANA
- Snowflake
- SAP BO
- PowerBI
- -
Business Process Improvement Analyst @ Medtronic
Prague, Czechia - On-siteSummary:
- Develop metrics for process measurement. Measure performance against process requirements. Perform cost and benefit analyses.
Responsibilities:
- Develop metrics for process measurement
- Measure performance against process requirements
- Perform cost and benefit analyses
- Implement process optimization initiatives
- SQL
- SAP ERP
- Snowflake
- Excel
- PowerBI
- -
Fraud Investigation Analyst @ Amazon
Prague, Czechia - On-siteSummary:
- Identify and minimize fraud risks in TR/GB markets. Communicate with customers, buyers, and merchants.
Responsibilities:
- Identify and minimize fraud risks in TR/GB markets
- Communicate with customers, buyers, and merchants
- Analyze transaction patterns and fraud indicators
- Implement risk mitigation strategies
- SQL
- Excel
Education
MSc Big Data & Artificial Intelligence
Thesis: Bridging Computational Foundations to Generative AI: A Design-Based Framework for Progressive Creative Coding Education
MA International Business & Economics
Thesis: The Role of Multinational Corporations in International Business: Rentier Capitalism
BSc International Trade & Finance
Projects
Data platform for Berlin's public transport — REST ingestion into PostgreSQL with live tracking and historical replay, served through FastAPI and Docker with a Leaflet.js map. On top of it, an LLM agent built from scratch: the raw tool-calling loop with no framework, an MCP server and client, an eval suite, Langfuse observability, and a LangGraph rebuild, running on local Ollama.
- REST ingestion into PostgreSQL with live tracking and historical replay
- Raw tool-calling loop with no framework, then rebuilt on LangGraph
- MCP server and client (FastMCP, read-only SQL) with retries and guardrails
- Eval suite and Langfuse tracing, running on local Ollama
- Python
- FastAPI
- MCP
- LangGraph
- Langfuse
- Ollama
- Docker
- PostgreSQL
- Leaflet.js
Open-source AI curriculum (15 modules) built as a Master's thesis, guiding learners from pixel manipulation to GANs, VAEs, and Diffusion Models. A user study with 9 participants showed knowledge scores rise from 15.1% to 50.5% (Cohen's d = 1.615, p = .008).
- 15 modules: pixels, geometry, classical ML, GANs, VAEs, Diffusion Models
- Pre/post test: 15.1% to 50.5% (Cohen's d = 1.615, p = .008)
- Dual-format pedagogy: hands-on discovery and conceptual deep-dives
- TouchDesigner integration for creative coding workflows
- Python
- Jupyter
- PyTorch
- TouchDesigner
- NumPy
- TensorFlow
- scikit-learn
- OpenCV
Real-time generative art installation combining boid flocking with an evolutionary algorithm — stars evade predator black holes and pass survival traits across generations. Exhibited at the Pixels2GenAI exhibition, Berlin (2026).
- Evolvable boid flocking (separation, alignment, cohesion)
- Predator black holes with gravitational lensing
- Tournament selection and uniform crossover across generations
- Quorum sensing: density-triggered collective defense
- Python
- Pygame
- NumPy
Real-time generative artwork where live GPU vitals — temperature, clock, utilization, power — sculpt a mycelial Physarum simulation coupled to a Hebbian neural network. Up to 200,000 agents render at ~60 FPS via GLSL compute shaders.
- Physarum agents sense, rotate, and deposit pheromone trails
- Hebbian synapses form on agent co-activation
- GPU vitals drive color, bloom, and step size
- GLSL compute shaders, 60 FPS at 1920×1080
- Python
- OpenGL
- GLSL
- Pygame
- SciPy
- CUDA
Interactive art installation where visitors watch themselves dissolve into noise and crystallize as anime artwork via Stable Diffusion and ControlNet. Exhibited at IT Studio Academis, Berlin (2026), with async generation at ~3s/frame and live display at 30 FPS.
- ControlNet conditioned on Canny edges and MediaPipe landmarks
- DDPM with linear, cosine, and quadratic noise schedules
- Async generation pipeline keeping display at 30 FPS
- Optional sound reactivity and ambient audio synthesis
- Python
- Stable Diffusion
- ControlNet
- MediaPipe
- OpenCV