About
I work where data, product, and AI meet. Shipping iOS apps, validating credit risk models, automating SAP master-data workflows, translating stakeholder needs into agile sprints. MSc in Statistics & Data Science (ML specialization) from LMU Munich, two and a half years at Infineon Technologies on SAP master data and workflow automation, a year as Product Owner at SCW.AI, and a first data job validating credit risk models at Yapı Kredi Bank. Alongside that I build and release my own AI-native iOS apps. Industrial Engineering foundation, across semiconductors, supply chain, and finance.
- Python
- SAP
- SQL
- Swift / iOS
- Agentic AI & RAG
- KNIME
- Power BI
- Agile / Scrum
Selected Projects
More projects
mispick
An open-source tool that measures which of an AI assistant's tools a model mixes up. It generates realistic test requests for every MCP tool, asks a real model to choose, and reports a confusion matrix with confidence intervals. It never runs a tool. Proposed description rewrites are re-tested and kept only if they measurably help. Also a GitHub Action, and the base of a pre-registered study.
Trames
An AI-generated narrative game for iOS. Players input a life decision and receive a personalized 14-day branching story; an agentic pipeline handles plot branching, character continuity, and daily delivery. Built solo in Swift.
Master's Thesis & hasa
"Robust SGD by Developing a Sample Selection Algorithm on Image Classification Tasks." Includes hasa, a Python package implementing the proposed sample-selection method, published on PyPI.
Agentic Survey Coder
An agentic AI pipeline built for a consulting engagement that automatically codes open-ended survey responses on relevancy, effort, and clarity (1–3 scale), replacing manual qualitative analysis.
Agentic RAG
A Python RAG service that answers questions over technical PDFs and returns answers with cited source pages. Combines vector search and BM25 with reciprocal rank fusion and cross-encoder reranking, plus an LLM-as-judge eval harness (15-question golden set) tracking retrieval hit-rate and answer faithfulness. Shipped as a FastAPI service in Docker.
Agentic Ticket Pipeline
An agentic incident-to-ticket pipeline for a chip-development and functional-safety setting. A Gemini agent reads a messy incident report, investigates it with tools over synthetic data, reasons about the root cause, and drafts a structured ticket. Around the agent sits a deterministic reliability layer: outputs must pass a strict Pydantic schema, the root cause must be grounded in the evidence the tools actually returned, and anything touching a safety-rated (ASIL / ISO 26262) component is routed to a human regardless of confidence. Built with FastAPI and streams the reasoning trace live.
Investment Analysis Dashboard
A Python dashboard that pulls portfolio data, runs price prediction models and adjusts strategy from the results. Built end to end with AI coding agents (Cursor), apart from the model training, which I did myself.
Experience
Mar 2024 — Sep 2026
Data Analyst, Working Student
Infineon Technologies · Munich, Germany · Master Data Department
- Built ETL and workflow automation in KNIME over SAP source systems, cutting manual processing time by roughly 40% (about 10 hours per week).
- Designed root-cause error-analysis procedures for SAP master data in SQL, reducing recurring data-quality errors across reviewed records by roughly 15%.
- Supported the rollout of DocJockey, an AI document indexing and review tool, contributing domain expertise on internal data workflows and document structures.
- Coordinated cross-functional teams on data migration to a unified SAP platform, keeping the work aligned with global data governance standards.
Aug 2022 — Sep 2023
Product Owner & Data Analyst
SCW.AI (Supply Chain Wizard) · İzmir, Turkey · Full-time
- Led product development by translating stakeholder requirements into user stories and technical specifications, cutting the mid-sprint rework and clarification cycles that used to delay delivery.
- Delivered data-driven insights using SQL, Python, and statistical methods that shaped product strategy, contributing to an estimated 20% increase in feature adoption.
- Managed client relationships and provided analytical consulting across 3+ key business accounts.
Sep 2021 — Jul 2022
Data Scientist Intern
Yapı Kredi Bank · Istanbul, Turkey · Risk Validation Department
- Validated and back-tested credit risk models, applying statistical tests that improved model accuracy (Gini) by ~3% and ensured regulatory compliance.
- Automated model monitoring reports using Python and SQL, reducing monthly reporting lead time by ~30%.
- Developed validation frameworks across 5+ risk models for robust performance evaluation.
Education
Graduated May 2026
M.Sc. Statistics & Data Science — Machine Learning Specialization
Ludwig-Maximilians-Universität München · Germany
Focus: Data-driven decision making, predictive analytics, statistical computing, deep learning.
Thesis: Robust SGD by Developing a Sample Selection Algorithm on Image Classification Tasks.
Aug 2017 — Aug 2022
B.Sc. Industrial Engineering
Bilkent University · Ankara, Turkey
Honors: Cum Laude (GPA 3.36/4.0). Strong foundation in optimization algorithms, stochastic processes, and system engineering. Ranked top ~0.15% in Turkish university entrance exam (~3,000 / 2M+ candidates).
2013 — 2017
High School Diploma · Science Track
İzmir Fen Lisesi · İzmir, Turkey
Graduated from one of Turkey's nationally-selective science high schools (Fen Lisesi system; admission top ~0.1% nationally).
Skills & Stack
Programming
- Python
- SQL
- Swift / iOS
- R
- Pandas
- NumPy
- Scikit-learn
- PyTorch
- TensorFlow
- Statsmodels
- FastAPI
- Docker
- Git
AI & Automation
- Agentic AI Pipelines
- LLM API Integration
- RAG Systems
- Hybrid Retrieval (Vector + BM25)
- LLM Evaluation
- LangChain
- LlamaIndex
- Prompt Engineering
- AI Coding Agents (Cursor, Claude Code)
- Process Automation
Enterprise & Data
- SAP
- KNIME
- ETL Pipelines
- Data Quality & Governance
- Power BI
- Tableau
- Metabase
- Excel (VBA)
- Jira
Statistics & ML
- Supervised & Unsupervised Learning
- Deep Learning
- Bayesian Inference
- Time Series Forecasting
- Hypothesis Testing
- A/B Testing
- Feature Engineering
- Regression
- Predictive Modeling
- Anomaly Detection
- Risk Model Validation
Product & Methods
- Agile / Scrum
- Product Ownership
- Stakeholder Communication
- KPI Development
- Dashboard Design
- Data Storytelling
- Requirements & User Stories
- Process Mapping
Certifications & Training
Aug 2026 — Sep 2026
Machine Learning Engineer – Associate Learning Plan
AWS Skill Builder
Hands-on Builder Labs covering data preparation, model training, deployment and monitoring on AWS. This is the learning plan with labs, not the AWS certification exam.
Dec 2025 — Mar 2026
RAG and Agentic AI Professional Certificate
IBM · Coursera
Retrieval-augmented generation, agent design and tool-augmented LLM applications. View certificate
Dec 2022 — May 2023
Machine Learning & Data Science Professional Certificates
IBM · Coursera
Supervised and unsupervised learning, deep learning and the applied data science workflow. View certificate
Languages & Beyond Work
Turkish
Native
English
C1 · Advanced
German
B1 · B2 course in progress
- Stand-up comedy (performing)
- Hiking
- Skiing
- Traveling
- Football
- Driving licence · Class B (DE)
Ask AI-Selim
AI-Selim · an AI trained on my CV. It can be wrong. For anything important, email me.
Hi, I'm an AI version of Selim. Ask me about my work, projects or studies.
Questions are processed by Google Gemini and not stored by me.