Artem Okhten

• Founded and lead EECAA, scaling membership to 50+ active members while establishing governance and operational structure
• Engineered and deployed a full-stack platform for automated news, events, and media management
• Built custom rich-text editor and implemented backend concurrency handling for reliable content workflows
• Designed system architecture including role-based access control, admin dashboards, and scalable workflows

Student Success and Support Center (SSSC), Florida Tech

• Tutored 300+ students in mathematics and physics, improving overall class averages by 20 - 30%.
• Supported SSSC's mission to improve STEM success rates by 25%.
• Designed math visuals and study materials adopted by multiple tutoring groups.
• Strengthened leadership and communication skills through peer mentoring and group instruction.

Darhaal Games Platform

Next.js, Supabase Realtime, WebSockets

• Engineered a real-time multiplayer system tested with 70+ concurrent users across multiple sessions
• Designed low-latency state synchronization using WebSockets
• Built extensible architecture supporting multiple multiplayer games (Battleship, Minesweeper)
• Focused on system reliability and smooth real-time user interaction

Chess Explainable AI

PyTorch, Deep Learning, Explainable AI

• Trained a CNN model on 10K+ chess games achieving ~60% prediction accuracy from mid-game positions
• Applied explainability methods (Grad-CAM, Integrated Gradients, Saliency Maps)
• Compared interpretability methods to evaluate explanation consistency and model behavior
• Investigated limitations of prediction in high-uncertainty environments

Statistical Learning Models

Python, Pandas, Scikit-Learn

• Built machine learning pipelines for classification and regression tasks on complex datasets
• Engineered and selected features to optimize cross-validation performance
• Evaluated statistical behavior of models to identify edge cases and error distributions

Chess Blunder Prediction

Python, Pandas, Scikit-Learn

• Developed a statistical model quantifying human error probabilities during complex board states
• Extracted and engineered features representing time pressure, board complexity, and evaluation deltas
• Processed and analyzed large-scale datasets to uncover behavioral decision-making patterns

USCIS Prep Assistant

React, TypeScript, PWA

• Built and deployed a production-ready SPA for U.S. citizenship test preparation (Jan 2026)
• Implemented fast search and optimized state management for instant response
• Designed a clean, mobile-first UI focused on accessibility and user retention