Egemen Eroglu

Hi, I'm Egemen, building & testing ideas that ship at Express speed.

R&D Specialist at DHL Express. AI training delivered to 200+ people · predictive ML systems running in production logistics.

Tap the van to honk.

Shipping status

Idea
R&D
Testing
Shipped

Selected Projects

ML · Last-MileIEEE UBMK 2025 ↗

Address Error Detection

Predictive ML framework that catches address errors before shipment, improving last-mile delivery success rates and reducing manual handling.

PythonTransformersXGBoostLightGBMCatBoost
Deep Learning · RoutingINFUS 2026 (accepted) ⏳

Destination Facility Prediction

Deep learning framework for predicting the correct destination facility for inbound shipments, reducing misrouting and handling costs.

PythonDeep LearningPandas
AI · LLM · MCPInternal 🔒

MCP Customs Portal

Hybrid RAG + LLM + MCP system for DHL Express Turkey customs operations, resolving unstructured regulatory queries (legislation PDFs) and structured EVRIM/BILGE data via natural language. Incorporates sentiment analysis, topic modeling, and hybrid keyword-semantic retrieval; targeting 90%+ intent classification accuracy, Recall@5 above 85%, and sub-5-second latency, with autonomous handling of 30% of routine call center queries. Patent application filed for the secure RAG + MCP integration component.

PythonMCPLLM
RPA · OCR · AutomationInternal 🔒

Freight Documentation Automation

Automated end-to-end generation of Freight Non-Existence Letters across 150-200 daily inbound shipment requests, cutting per-request processing time from 1.5 minutes to under 15 seconds via a Snowflake → Dataiku → MSSQL → IIS pipeline with full audit logging, freeing 0.5 FTE and extending availability from business hours to 7/24.

PythonRPAOCR
OR · EV InfrastructureISPR 2026 (accepted) ⏳

EV Charging Station Optimization

Formulated a binary integer linear program for EV fleet allocation across DHL Express Turkey's 3 Istanbul distribution centers, jointly optimizing customer-to-depot and vehicle-to-depot assignment to minimize demand-weighted total distance under capacity constraints. Solved via PuLP/CBC.

PythonILPPuLP/CBC
NLP · MLOps · ITSMInternal 🔒

ARGUS — ITSM Intelligence Platform

Built an AI-powered ticket intelligence system for DHL Express Turkey's IT support team, combining a fine-tuned Turkish BERT classifier, a LightGBM SLA-risk scorer with SHAP explainability, and a FAISS-based similar-ticket retrieval engine over historical ServiceNow data. Served through a single FastAPI endpoint with a web UI supporting drag-and-drop ticket parsing and analysis history, to speed up IT operations triage and routing.

PythonBERTLightGBMFAISSFastAPI

About

I'm an R&D Specialist at DHL Express: researching, prototyping, and deploying ML systems that run in DHL Express's global logistics network.

FocusThe model is 20% of the work. The other 80% is keeping it alive.
FunStrong coffee · mechanical watches · saxophone.
LanguagesTurkish (native) · English (C1) · German (A2)
LocationIstanbul, open to relocation in Germany.

ML & AI

PyTorchscikit-learnTransformersXGBoostLightGBMCatBoostOptunaSHAPLIMELLM/RAG pipelinesAgent Orchestration (MCP)

Languages

PythonSQLUnix/Linux ScriptingGit

Cloud & Databases

Azure (AZ-900, DP-900)SnowflakeDatabricksMSSQLPostgreSQLMongoDBTeradata

Data Engineering

AirflowKafkaDockerSparkDataikuIBM DataStage

Frameworks

DjangoFastAPIStreamlitPydanticPytest

BI & Monitoring

PowerBITableauGrafanaELK Stack

Experience

Experience

🚚

R&D Specialist · DHL Express

Apr 2025 – Present

Patent application filed · AI training for 200+ · DigiForce · RPA leadership

🏭

Data Scientist · Bosch

Oct 2023 – Mar 2025

AI-SCREP · SparkUp / GecoAI

🧠

ML Researcher · Koç University

Jul 2023 – Jul 2024
🏦

DWH & Data Engineer · Halkbank

Jul 2023 – Oct 2023
💻

Data Engineer Team Lead · Jeton Digital

Mar 2023 – Jun 2023
💻

Data Engineer · Jeton Digital

Sep 2022 – Mar 2023
📊

Data Scientist Intern · SmartOpt

Jul 2022 – Aug 2022

Education

🎓

MSc Computer Engineering · Bahçeşehir University

Oct 2025 – Present
🎓

BSc Computer Engineering · Bahçeşehir University

Sep 2020 – Jan 2025

Publications

📄

From Detection to Resolution: A Deep Learning Framework for Destination Facility Prediction in Inbound Shipments

INFUS 2026
📄

Post-Hoc Explainability for Neural Anomaly Detection in Logistics: An Empirical Comparison of SHAP, LIME, and Integrated Gradients

UBMK 2026
📄

Set Covering Problem-Based Location Optimization for Electric Vehicle Charging Stations: A Real Case Study

ISPR 2026
📄

AI and Operations Research Applications in Express Air Cargo: A Systematic Literature Review

YAEM 2026

Contact

Let's talk

Best way: erogluegemen@gmail.com

Also on: GitHub · LinkedIn · Google Scholar

Resume

Updated July 2026.

View Resume ↗