About

Machine Learning Engineer with a strong mathematical background, experienced in building end-to-end ML systems in PyTorch. Worked on research-driven generative models for structured and relational data, including sequence- and network-based representations, covering data processing, model design, training, and evaluation.

Work Experience

M&C Saatchi World Services
Python
TypeScript
FastAPI
NetworkX
Pyvis

Jul 2025 - Aug 2025

Data Intern

  • ⦿ Developed a time-series network visualization and prediction tool using state-of-the-art DAMNETS model, enabling stakeholders to identify emerging influence patterns and forecast network evolution across social media data.
  • ⦿ Presented findings to non-technical stakeholders across multiple teams, driving adoption of network analysis for client account strategy.
  • ⦿ Top-performing intern; recommended for return offer by management and data science team.

Thesis

  • ⦿ A. Tuna, "A Deep Autoregressive Model for Dynamic Combinatorial Complexes," arXiv preprint, 2026. DOI: 10.48550/arXiv.2503.01999.

Education

Imperial College London

Oct 2024 - Oct 2025
Master of Research (MRes) in Artificial Intelligence & Machine Learning
2:1

University of Edinburgh

Sep 2019 - May 2024
Bachelor of Science (BSc) in Mathematics & Statistics
2:1; transferred from Mechanical Engineering in second year.

Projects & Awards

Developed a novel autoregressive generative model for structured, high-dimensional graph and combinatorial data (arXiv:2503.01999). Implemented the model in PyTorch with custom loss functions, evaluation pipelines, and memory-aware handling of sparse, high-dimensional structures.

PyTorch
Scikit-Learn
NetworkX
TopoModelX
Torch-Geometric

Pabu

3rd Place

Co-designed and implemented a Chrome extension for AI-driven content filtering, selectively blocking distracting online material while allowing context-relevant resources. The project was developed collaboratively during the AI Engine Hackathon and awarded 3rd Place for Best Use of ACI.dev Unified MCP.

LangChain
LangGraph
Pydantic
React
FastAPI
Node.js

HYPED

UK Winner

Designed and oversaw manufacture of the monorail–pod structural interface, enabling stable operation at speeds up to 311 km/h, for the UK’s leading Hyperloop team; Uthe project was recognised by SpaceX, Virgin Hyperloop One, and the Institution of Civil Engineers, and reached SpaceX finals in Los Angeles.

SolidWorks
MATLAB
COMSOL

ML Research & Modelling Skills

Python
PyTorch
Deep Learning
Generative Models
Graph Neural Networks
FastAPI
Docker
Git
Statistical Modeling
Network Science
Linear Algebra

Languages

English (Fluent)
Turkish (Native)
German (B2)
French (A1)

Leadership

University of Edinburgh, School of Mathematics

Sep 2022 - Aug 2023

MathClans Leader

  • ⦿ Fostered a sense of community by leading peer integration initiatives for new students.
  • ⦿ Organised academic and social events supporting student engagement and departmental cohesion.

Edinburgh University Students' Association

Oct 2020 - Aug 2023

Programme Representative, Academic Societies Representative

  • ⦿ Gathered feedback and represented third-year Mathematics students in the Student-Staff Liaison Committee.
  • ⦿ Liaised between academic societies and the Association, contributing to governance via motions and proposals.