Open to opportunities

Kartikeya User

@kartikeyauser

Message

Machine Learning researcher specializing in adversarial robustness and fraud detection.

Germany

Message

What I'm looking for

I seek research/engineering roles focusing on adversarial robustness, fraud detection, and scalable ML systems, with collaborative teams and applied impact.

I am an experienced Machine Learning researcher with 3+ years building production-ready AI models and research systems focused on robustness and adversarial evaluation. My work bridges deep learning, multimodal models, and practical deployment for real-world applications.

At Zuse Institute Berlin and in internships, I implemented adversarial attacks, supervised fine-tuning, and comprehensive evaluation pipelines using PyTorch, Transformers and multi-GPU SLURM clusters. I have applied these techniques to improve model robustness and reduce spurious correlations in vision-language systems.

I developed end-to-end systems including a RAG-based product recommender using LangChain, OpenAI APIs, FAISS, and containerized deployments on Kubernetes and GCP. I also built predictive models with 85%+ accuracy using Scikit-Learn, Pandas and NumPy to support data-driven business decisions.

I communicate technical findings clearly to diverse stakeholders and collaborate with institutions on applied research in fraud detection and rare pattern discovery, combining strong quantitative skills, programming expertise in Python, and hands-on experience with LLMs and multimodal architectures.

Experience

Work history, roles, and key accomplishments

Student Research Assistant

Zuse Institute Berlin

Sep 2023 - Sep 2025 (2 years)

Evaluated adversarial robustness and explainability of vision-language models, implemented supervised fine-tuning and adversarial/counterfactual training, and developed multi-GPU evaluation pipelines that improved multimodal robustness assessment.

PyTorch Transformers Slurm Model Evaluation Counterfactual Training

Machine Learning Intern

TestAIng.com

Jun 2021 - Aug 2021 (2 months)

Implemented multiple adversarial attack algorithms on deep neural networks and produced reproducible documentation for model robustness experiments to support research and testing workflows.

TensorFlow Deep Learning Documentation Python

Data Analyst Intern

Peacock Solar

May 2020 - Jul 2020 (2 months)

Built predictive ML models and preprocessing pipelines to forecast market penetration and revenue across 28 states, producing models with 85%+ accuracy that informed expansion decisions.

scikit learn Pandas NumPy A B Testing data preprocessing Web Scraping Forecasting Python

Education

Degrees, certifications, and relevant coursework

Technische Universität Berlin

Master of Science, Scientific Computing

2022 - 2025

MSc in Scientific Computing with focus on numerical linear algebra, optimization, and machine learning; thesis on robustness of multi-modal foundation models including adversarial evaluation and sparse attacks.

University of Delhi

Bachelor of Science, Physical Sciences

2018 - 2021

BSc in Physical Sciences focused on numerical methods, calculus, differential equations, programming, and system architecture for computational applications.