About Me

About Maxim Podgore

Hi! I'm Maxim Podgore, a Machine Learning Engineer and AI Researcher based at the University of California, San Diego (UCSD). I currently serve as the Lead Student ML Engineer at UCSD ITS, where I lead initiatives in machine learning development and AI research.

Education

University of California, San Diego
Bachelor of Science in Computer Science with a specialization in Machine Learning and Data Science
GPA: 3.8/4.0 (expected graduation)

Technical Expertise

My technical skills span a wide range of modern technologies and frameworks:

  • Languages: Python, C++, JavaScript/TypeScript, LLVM, SQL
  • ML/AI Frameworks: PyTorch, TensorFlow, Scikit-learn, Hugging Face Transformers
  • Web Development: Next.js, React, Node.js, Tailwind CSS, Astro
  • Tools & Platforms: Redis, Pinecone, Vercel, Git, Docker
  • Specializations: LLM development, Deep Kernel Learning, Bayesian Optimization, Reinforcement Learning, RAG systems

Research & Professional Experience

Lead Student ML Engineer - UCSD ITS
I lead machine learning initiatives and AI research projects, working on cutting-edge solutions that push the boundaries of what's possible with modern AI technologies. My work involves extensive experience with large language models, implementing and extending LLM functionality for real-world applications.

Notable Projects

dReal-clang-tidy - Developed a sophisticated C++ static analysis tool that uses graph-based solvers over LLVM Abstract Syntax Trees to validate floating-point rounding mode preconditions, detect contradictions, and generate structured JSON outputs. This project combines compiler technology with formal verification methods.

RippleEdits Benchmark Extension - Extended the RippleEdits benchmark for SingularityNET's graphRAG system, creating comprehensive evaluations of knowledge editing effects in Retrieval-Augmented Generation systems. This work contributes to understanding how knowledge modifications propagate through graph-based RAG architectures.

Groundwork Books Full-Stack Platform - Built a production-ready website using Next.js, implementing Redis for efficient caching, Pinecone for semantic search capabilities, and modern Tailwind styling. The platform is hosted on Vercel and serves as a complete e-commerce solution for a local bookstore.

TuRBO-O Algorithm Enhancement - Improved the Turbo Algorithm by integrating a global kernel updated via Deep Kernel Learning and incorporating the Upper Confidence Bound Algorithm for enhanced exploration/exploitation balance in Bayesian optimization tasks.

Blackjack AI - Developed a comprehensive reinforcement learning agent using Monte Carlo methods, Temporal Difference learning, and Q-Learning. The system learns optimal blackjack strategies through self-play and features real-time visualizations of state values and decision-making processes.

Achievements & Recognition

  • GitHub Pull Shark x2 Achievement
  • Arctic Code Vault Contributor
  • YOLO & Quickdraw Achievements
  • 8 followers and active open-source contributor with 33+ repositories
  • Multiple starred projects including Groundwork Books (5 stars), TuRBO-O (2 stars), and AztecPlanner (2 stars)

Community & Collaboration

I'm passionate about open-source development and collaborative projects. I've contributed to various initiatives ranging from hackathon projects like AztecPlanner (ACM SDSU Innovate 4 Hackathon 2025) to research-oriented repositories. I believe in the power of community-driven development and am always eager to work on innovative projects that push technological boundaries.

Let's Connect

I'm always interested in collaborating on new projects, particularly those involving machine learning, AI research, or innovative web applications. Feel free to reach out through: