About Me
Software engineer at Amazon's Finance Automation org in Seattle, WA. Working on using AI to automate various areas inside Amazon's finance division. My technical interests include machine learning (LLMs, RAG, test-time inference, training and fine-tuning models), full-stack development, and designing scalable systems.
Education
I graduated from UC Berkeley with an honors degree in Electrical Engineering and Computer Science in May 2024. I took numerous grad-level and upper-division courses on machine learning, optimization, probability, databases, operating systems, and more.
Research Experience
I've worked at some of the most prestigious research institutions in the world. As a Machine Learning Researcher at Lawrence Livermore National Laboratory, I worked in the field of quantum physics — modelling particle collision data to predict anomalous events in future data, learning in depth about the GAN architecture and its applications. At Wheeler Labs at Berkeley, I applied NLP-based text classification across qualitative plans related to Groundwater Sustainability, utilizing LLM outputs to derive actionable recommendations. I also interned at Bell Labs in summer 2022, researching "advertisable" products on Web3.
Current Role
Software Development Engineer at Amazon's Finance Automation org. Using AI to automate various areas inside Amazon's finance division. Tech stack: AWS, Java, TypeScript, Node.js, and Python.
Projects
I like creating projects for fun. Currently building an intelligent system to store and classify large amounts of financial data, retrieve relevant context per query efficiently, and reason/predict based on current context and market trends — with self-adaptive weights at inference time. My projects range from interactive games to web apps to training and optimizing language models.
Resume
Contact
Personal: alpaltug@berkeley.edu
Work: alpaltug@amazon.com