I'm a software engineer transitioning into independent AI research, focused on LLM reliability and neural network optimization. After several years building production mobile apps and deploying ML systems (computer vision, NLP), I shifted toward formal research — co-authoring two preprints on principled optimizer design and systematic characterization of LLM failure patterns. I'm currently applying for thesis-based MSc programs in Canada to continue this work.
👀 More Details
- 🔬 I'm currently doing independent AI research on LLM reliability and optimizer design — see my publications below.
- 💼 My professional background is in software engineering (React Native, production ML deployment).
- 👨💻 All of my public projects are available on this page.
- 💬 Ask me about LLM evaluation, optimizer design, statistical methodology, or React Native.
- 📫 The fastest way to contact me is via my email.
📄 Publications
-
Emergent Brittleness: A Case Study on Pathological Instruction-Following in LLM-Based Code Optimization
S. Tatlari*, S. Mobarhan*, H. R. Hamidi. Preprint, 2025.
📄 DOI | 💻 Code -
Unlocking Potential: A Case for Neural Network Perseverance on Critical Tabular Tasks Through Principled Optimizer Design
S. Mobarhan*, S. Tatlari*. Preprint, 2025.
📄 DOI
*Denotes equal contribution.
