All Projects

Explore all my projects in a filterable database

Showing 9 of 9 projects

Will it Blend?

In Progress

October 2024 - Present

High-performance automated Blender render system with GPU acceleration, automatic video compilation, crash recovery, and cloud sync. Features CUDA optimization, automatic EXR-to-MP4 compilation with ACES color support, checkpoint-based crash recovery, and Google Drive integration.

ENGINEERING PythonBlenderCUDAFFmpegGoogle Drive API

JAMES (Automated Lab Book Generator)

On Hold

January 2025 - December 2025

Architected multimodal system generating lab books from recorded lab sessions using speech recognition, speaker diarization, and LLMs with RAG knowledge base for contextual understanding across experiments. Designed modular architecture supporting both local LLMs (Ollama) and cloud APIs (OpenAI/Anthropic) for flexible deployment.

AI & ML PythonWhisperPyTorchLLM

AgriView - LauzHack

On Hold

November 2024 - December 2025

Participated in the LauzHack hackathon in Lausanne, developing an application that uses NLP interfacing with a computer vision image segmentation model (PASTIS) to predict crop yield based on open source satellite imagery.

AI & ML Python

Taming the Tokamak (Final-Year Project)

Completed

2025 - 2026

"Taming the Tokamak: MHD Instabilities and Control Strategies towards Burning Plasmas" - final-year group project supervised by Prof. Stuart Mangles (78%). Reviewed sawtooth crashes, neoclassical tearing modes and resistive wall modes, assessing the readiness of ECCD, ICRH and active feedback control for ITER.

PHYSICS MHDTokamak PhysicsLiterature Review

Physics Informed Deep Learning

Completed

May 2024 - July 2024

Developed physics-informed neural networks (PINNs) to identify the governing equations of non-linear dynamical systems, comparing the efficacy of different neural network architectures and hyperparameters. Worked with Long Short-Term Memory networks and Reinforcement Learning Ran training on the Imperial High-Performance Cluster Achieved 75% overall score

PHYSICS PythonGitPyTorchTensorFlowMATLAB

Sentiment Analysis Trading Bot

Completed

June 2024 - December 2024

Used Alpaca API to get stock data and news articles. Implemented a transformer and Lumibot to execute trades every hour based on sentiment analysis of news articles. 20.7% annual return on investment in backtesting.

AI & ML PythonAPIPyTorch

Interdisciplinary Research Computing

Completed

January 2025 - March 2025

Runner-up, Best Project - used OpenFOAM to model the effect of various pathologies on blood flow in the human aorta. Developed computational fluid dynamics (CFD) simulations with varying degrees of stenosis (0-80% occlusion) to study blood-flow patterns. Created parametric simulations to analyse wall shear stress, pressure gradients, and flow velocities in stenotic aortas.

PHYSICS PythonOpenFOAM

Analogue Computing

Completed

January 2024 - May 2024

Designed and built an analogue computer circuit for solving second-order differential equations Created computational simulations of damped oscillatory systems to model real-world car suspension behaviour Achieved 83% overall score

ENGINEERING PythonLTspiceSciPy

Robot Arm Control System

Completed

November 2025 - January 2026

Programmed a microcontroller to control a robot arm with all control code written in assembly language Implemented motor control algorithms and real-time feedback systems for accurate positioning and movement

ENGINEERING AssemblyMicroprocessor

© 2026 Lucas Ertugrul. Built with SvelteKit & Tailwind CSS. Unashamedly vibe coded.