Girish Kumar

I'm currently exploring Data Analyst, Data Scientist and ML Engineer roles where I can apply deep learning to meaningful problems. Open to opportunities in Berlin or anywhere in Germany or remote.

Independent Data Scientist / ML Engineer December 2024 – Present
Self-Employed Rostock, Germany

• Developed end-to-end ML pipelines from data preprocessing to model deployment, following MLOps best practices
with MLflow for experiment tracking and model versioning.
• Built SkinSavant, an AI-powered diagnostic platform using Deep Learning with DenseNet169 achieving 90+ percent
accuracy, demonstrating Proof-of-Value model development and validation.
• Integrated Generative AI including GPT, LangChain, and RAG systems for intelligent conversational interfaces,
showcasing modern AI approaches.
• Conducted exploratory data analysis and feature engineering on diverse datasets, translating business questions
into data-driven solutions.
• Applied transfer learning and modern algorithms including EfficientNetB0 and MobileNetV2 for scalable,
production-ready solutions.


Data Scientist / Data Analyst July 2022 – November 2024
Digital Spine GmbH Rostock, Germany

• Developed and deployed ML models on Databricks with MLOps practices including MLflow and CI/CD, improving
forecasting accuracy by 25 percent.
• Translated complex business requirements into clear, data-based questions and delivered measurable results for
stakeholders.
• Built scalable data pipelines using Python and SQL, integrating multiple data sources into the analytics
architecture.
• Collaborated in agile teams with Product, Engineering, and Operations, contributing to the Data Science community.
• Continuously optimized existing models and documented solutions following MLOps frameworks.
• Created dashboards in Power BI and Tableau, communicating insights to technical and non-technical stakeholders.

Data Analyst February 2019 – July 2022
CENTOGENE GmbH Rostock, Germany

• Applied ML classification algorithms on large-scale clinical datasets, improving anomaly detection accuracy by 25
percent.
• Identified new data sources and conducted exploratory analyses to uncover business insights.
• Developed SQL queries optimizing data extraction efficiency by 40 percent, enabling data-driven decision making.
• Partnered with scientists and engineers to develop data-driven solutions for clinical diagnostics.

Data Analyst Intern May 2018 – Feb 2019
Unicomp Mittweida, Germany

• Improved data retrieval times by 15%, as measured by average query runtime, by developing optimized
SQL queries for large customer and transactional datasets supporting marketing analytics initiatives
• Enhanced report clarity and stakeholder accessibility by 25%, as measured by user feedback scores,
by creating interactive dashboards in Tableau and Power BI that visualized customer acquisition and
conversion metrics

Ironhack – Data Science and Machine Learning Bootcamp September 2025 – December 2025
Deep Learning, NLP, Computer Vision, Generative AI, RAG Systems, LangChain, MLOps, Agile Methods
Hochschule Mittweida – Master of Science 2016 – 2019