I'm Atakan as in Safak Atakan Celik
- 🔭 Currently working on machine learning and business intelligence as a data analyst.
- 💡 Exceptionally interested in statistics, languages, NLP, communications, neural networks, brains, nervous systems.
- 🤖 Supporting a collaborative future of humans and machines with realistic steps.
- 🌱 Always be learning.
- 👯 Looking to collaborate on side projects to make education accessible.
- 📫 To reach me use: www.linkedin.com/in/safakatakancelik or safakatakancelik@gmail.com
- ❤️ Pronouns: he/they/them
- 👾 Curious about: Future of VR, AR, and how things would happen if we were to discover new intelligent species?
| Project | Description | Tech Stack / Tags | Links |
|---|---|---|---|
| Collaborative Data Framework | Designed repository structure to boost agentic data science workflows using self-contained project workspaces. | Agentic Workflows, Data Science, Repository Architecture, Automation |
GitHub |
| RAG & Evaluations | Autonomous RAG system evaluated with the RAGAS framework on the HotPotQA dataset for grounding analysis. | RAG, RAGAS, HotPotQA, Query Expansion, LLM-as-a-Judge |
GitHub |
| Vector DB Explorer | Vector database engine & GUI for document ingestion, chunking, embedding, vector plotting, and hybrid search. | ChromaDB, Chunking, Embedding, Similarity Search, Visualization |
GitHub |
| Google 5-Day AI Agents Intensive Capstone: Onboarding Agent | Conversational agentic onboarding system replacing legacy forms by executing API tool calls dynamically. | Google ADK, Python, Flask, PostgreSQL, Supabase, Sheets API |
GitHub • Video |
| Microsoft AgentCon - Pizza Ordering Agent Workshop | Autonomous pizza-ordering agent using RAG, tool calling, and MCP server integrations over a live API endpoint. | Agents, RAG, Tool Calling, MCP Servers, Azure, Foundry |
GitHub |
| Project | Description | Tech Stack / Tags | Links |
|---|---|---|---|
| Neural Network from Scratch | Multilayer perceptron built using pure NumPy for classification with manual forward/backpropagation. | Neural Networks, Backpropagation, Gradient Descent, NumPy |
Notebook • Video |
| Time Series Forecasting (Mobi Bikes) | Production-ready SARIMA model isolating complex seasonal patterns to forecast bike trip demand. | Time Series, SARIMA, Python, Databricks, pmdarima, statsmodels |
Notebook |
| Customer Segmentation & Strategy | RFM analysis and K-Means clustering model for categorizing customer purchasing personas. | Customer Segmentation, RFM Analysis, K-Means, EDA, Feature Eng. |
Notebook |
| Mall Customers Segmentation | Full unsupervised ML pipeline (EDA, outlier filtering, scaling, elbow & silhouette analysis) for persona discovery. | Python, Scikit-Learn, K-Means, Clustering, Pandas |
Notebook |
| Attention From NumPy to PyTorch | Scaled dot-product attention built from scratch, covering QKV matrix theory, causal masking, and LLM decoding. | Attention, Transformer, PyTorch, Causal Masking, NumPy |
Notebook |
| Project | Description | Tech Stack / Tags | Links |
|---|---|---|---|
| Tableau Dashboard Stories: Profitability | Interactive dashboards and data stories analyzing 4-year KPI trends and product sub-category profitabilities. | Tableau, Dashboarding, Data Visualization, Data Storytelling |
4-Year Overview • Sub-Category Analysis |
| Tableau Dashboard: Profitability Overview | Executive dashboard tracking Superstore orders with dynamic filters and visuals for decision-making. | Tableau, Dashboarding, Data Visualization |
Dashboard |
| Looker Dashboard: Customer Churn | Interactive Looker visual analytics dashboard representing customer churn segmentation results. | Looker, Dashboarding, Customer Segmentation |
Project Writeup |
| Project | Description | Tech Stack / Tags | Links |
|---|---|---|---|
| Cloud Data Movements & Pipeline | End-to-end cloud data pipeline orchestrating data movement across Azure Blob Storage, CosmosDB, and Power BI. | Azure Data Factory, CosmosDB, Blob Storage, Power BI, Fivetran |
Project Writeup |
| SQL Data Pull Strategy | OOP Strategy Pattern implementation allowing dynamic data extraction across multiple SQL database types. | Python, SQL, Strategy Pattern, OOP |
GitHub |
| Cloud Gourmet DB Design | Relational schema design for a meal kit subscription platform (users, plans, deliveries, discounts). | SQL, Relational DB Design, pgAdmin |
ER Diagram |
| Book Cafe DB Design | Database architecture managing book inventory, orders, customer accounts, and supplier contracts. | SQL, Relational DB Design, pgAdmin |
ER Diagram |
| Healthy Life Clinic DB Design | Schema design managing patient records, medical practitioner schedules, and clinic appointments. | SQL, Relational DB Design, pgAdmin |
ER Diagram |
