Open to opportunities

Dilan Üzümcü

@dilanzmc

Message

Data Scientist turning multimodal data into measurable business impact through production-grade ML systems.

Turkey

Message

What I'm looking for

I’m looking for a production-first team where I can own end-to-end ML—from data ingestion and evaluation to deployment—especially in multimodal and data-pipeline-heavy roles, with room to scale systems using rigorous, measurable experimentation.

I’m a Data Scientist with 3+ years of end-to-end experience designing and deploying production-grade AI and data systems—from ingestion and feature engineering to modeling, evaluation, and deployment. My work centers on turning complex, heterogeneous data into measurable outcomes, grounded in strong statistical foundations and a production-first engineering mindset.

I’ve built production multimodal pipelines for large-scale video topic labeling and automatic labeling (using KNN with vision-language and large language models), and architected an AI-powered Geo-Spatial Business Intelligence & Recommendation platform that unifies multi-source data into a data lakehouse with data-quality automation. I also develop weakly supervised segmentation and geo-spatial capture-rate models, and automate the full ML lifecycle with AWS SageMaker, S3, and MLflow. Alongside engineering, I mentor and teach data-science workflows, delivering hands-on ML training and statistical consulting.

Experience

Work history, roles, and key accomplishments

Current

Data Scientist

Current

Bilgeadam Teknoloji

May 2026 - Present (2 months)

Built a production-grade multimodal video understanding and topic labeling system combining KNN nearest-neighbor retrieval with vision-language models and LLMs. Designed an end-to-end multimodal feature pipeline (embeddings, speech-to-text, OCR, object detection), evaluation framework, and taxonomy mapping to improve labeling quality and throughput.

Multimodal AI K NN Vision Language Models Large Language Models Speech to Text Object Detection Video Understanding Benchmarking Learning Curve Analysis Cross Validation Taxonomy Design Batch Processing

Data Scientist

Mapin Data

Jul 2023 - Feb 2026 (2 years 7 months)

Architected an end-to-end geo-spatial business intelligence and recommendation platform using heterogeneous data sources integrated into a data lakehouse. Built production ML pipelines including VLM-to-LLM structured feature extraction, weakly supervised multimodal segmentation, geo-spatial footfall/capture-rate modeling, and multi-task outputs deployed via AWS SageMaker with S3 and MLflow.

Special Educator & Stat Consultant

Saksi Kampus

May 2024 - Nov 2025 (1 year 6 months)

Designed and delivered hands-on ML training covering the full data-science workflow from problem framing and EDA through feature engineering, modeling, and evaluation. Provided statistical consulting and methodological guidance for interdisciplinary research, mentoring participants through end-to-end ML pipelines to deployment.

Statistical Consulting Mentoring Model Evaluation Machine Learning Training EDA Feature Engineering Modeling Evaluation Python Pandas NumPy scikit learn Matplotlib SeaBorn Plotly Scipy

Statistician / Data Analyst

Freelance

Jun 2022 - Aug 2022 (2 months)

Provided statistical modeling and analysis for research and business clients, including experimental design and hypothesis testing. Produced regression-based analyses and structured reporting using R and Python.