I am a Computer Engineering undergraduate and researcher specializing in Generative AI and Autonomous Systems. Currently, I serve as an Undergraduate Researcher at TÜBİTAK, investigating high-fidelity histopathology image generation using GANs and diffusion models.
Simultaneously, I work as an AI Engineer at Hagia Labs, where I develop agentic workflows and RAG pipelines to automate complex enterprise processes. My background is deeply rooted in robotics and embedded systems; I previously led a 30 member engineering team at Sema Aviation, deploying real-time perception stacks on NVIDIA Jetson hardware.
#Publications
- Code2Doc: A Quality-First Curated Dataset for Code Documentation↗
Recep Kaan Karaman, Meftun Akarsu•2025•arXiv:2512.18748 [cs.SE]
- RAG-Driven Data Quality Governance for Enterprise ERP Systems↗
Sedat Bin Vedat, Enes Kutay Yarkan, Meftun Akarsu, Recep Kaan Karaman, Arda Sar, Çağrı Çelikbilek, Savaş Saygılı•2025•arXiv:2511.16700 [cs.DB]
#Experience
TÜBİTAK
December 2025 — PresentUndergraduate Researcher (2247-C STAR)Investigating generative modeling approaches (GANs, diffusion models) for high-fidelity synthetic histopathology image generation. Building reproducible end-to-end experimentation workflows.
Hagia Labs
October 2025 — PresentAI Engineer Working StudentDeveloping agentic workflows and RAG systems for internal automation. Improved query accuracy by 38% and reduced manual review times significantly through AI-powered process efficiency.
Sema Aviation
March 2025 — December 2025Team Lead — AI EngineerLed a 30-member interdisciplinary team to design autonomous UAVs. Implemented real-time object tracking (YOLO-Seg + MobileSAM) and fault detection pipelines on NVIDIA Jetson hardware.
#Projects
Python Documentation RAG Assistantsource ↗
A lightweight RAG assistant for Python docs using semantic chunking, transformer embeddings, and cross-encoder reranking. Achieved 77% top-3 accuracy and visualizes the pipeline via Streamlit.
Cortex — CNN from Scratchsource ↗
A modular CNN framework built entirely in pure NumPy. Implements backpropagation, convolution, pooling, and optimizers from first principles, achieving 89% accuracy on MNIST.
Autonomous UAV System
Real-time perception stack using ROS 2, RF-DETR, and Kalman filtering. Integrated sensor fusion and control pipelines for autonomous flight on embedded NVIDIA Jetson platforms.
#Blog
Check Medium for my latest writing.
#Recommended Reading
These are papers, references, and books I revisit often — for their foundational insights, theoretical clarity, or relevance to my work in AI engineering, robotics, and systems design.
I also spend time listening to and learning from researchers whose work and thinking have shaped the field. I regularly follow talks and writing by Ilya Sutskever, and I closely track the research of others like Transformer Explainer, Neural Networks and Deep Learning (Michael Nielsen), Chip Huyen's Blog, Lilian Weng's Blog.
Foundational Papers
- Attention Is All You Need (Vaswani et al., 2017) — the original Transformer paper, foundational for sequence modeling and modern deep learning.
- Adam: A Method for Stochastic Optimization (Kingma & Ba, 2014) — widely-used optimizer with adaptive moment estimates; practically indispensable.
Practical References
- PyTorch Cheat Sheet — a quick reference I still glance at during prototyping.
- Transformer Explainer (Polo Club of Data Science) — an interactive visualization to learn how the Transformer model works.
- Visual Information Theory (Chris Olah) — a visual explanation of entropy, KL divergence, and information theory concepts.
- How to Use t-SNE Effectively (Distill.pub) — a hands-on guide to interpreting and using t-SNE responsibly.
- Sebastian Raschka’s Deep Learning Model Implementations — minimal PyTorch examples of ResNet, VGG, GANs, and others.
- A Recipe for Training Neural Networks (Karpathy) — a distilled set of practical tips and pitfalls for neural net training.
- PyTorch Official Examples Repository — solid, idiomatic examples of classic models and training pipelines.
Books I Recommend
- C++ Primer by Stanley B. Lippman, Josée Lajoie, and Barbara E. Moo — a thorough and well-paced guide to modern C++, ideal for building foundational systems knowledge.
- Designing Machine Learning Systems by Chip Huyen — great for connecting ML theory to real-world system design, especially in production pipelines.
- Mathematics for Machine Learning by Marc Peter Deisenroth, A. Aldo Faisal, and Cheng Soon Ong — covers linear algebra, calculus, probability, and optimization with clean motivation for ML applications.
- Introduction to Linear Algebra by Gilbert Strang — a foundational text for understanding vector spaces, matrix operations, and the structure of linear systems; highly recommended for building ML intuition.
- Neural Networks and Deep Learning by Michael Nielsen — a clear, code-first introduction to the core concepts of neural networks and deep learning.