.......

Diar Shakimov  • Peerlist

- Built a real-time General Relativity simulator using OpenGL and GLSL shaders to visualize Einstein’s spacetime curvature, creating an interactive educational tool demonstrating complex physics concepts. – Programmed dynamic 3D grid deformation algorithms demonstrating how massive objects bend the spacetime fabric with real-time gravitational field visualization. – Integrated advanced physics calculations, including Schwarzschild radius computations to ensure realistic gravitational effects.

Einstein's Spacetime Curvature Simulator

– Engineered iOS event management platform compiling university tech events from email and Discord servers using multi-source API integration, implementing date/time extraction to parse unstructured announcements into a unified event database. – Architected automated event scraping pipeline processing Discord server messages and email invites, extracting structured event data and synchronizing bidirectionally with Google Calendar API for seamless cross-platform accessibility. – Developed at StormHacks 2025, expanding scope to centralize all SFU CS-related events and branch into surrounding community events, addressing campus-wide event discovery challenges - Continued development and published to App Store

Event Stack (Tech Event Platform Published on

- Built a multi-class fraud detection system processing 99% imbalanced transaction dataset, engineering 5+ features including balance mismatch flags and error signals to capture data inconsistencies and suspicious patterns. - Implemented ensemble Random Forest classifier with class weighting and hyperparameter tuning (tree depth, splits, samples), optimizing for Macro F1-score to achieve equal performance across 4 urgency classes and winning 1st place

Fraud Detection Algorithm (1st Place SFU ML H

Built a tokenization-based static analysis engine in C++, detecting performance anti-patterns by tracking scope, variable lifetimes, and token context, producing JSON diagnostics with no external runtime dependencies. – Implemented a header-only micro-benchmark harness with IQR outlier removal and p95/p99 tail-latency tracking, measured up to 15× throughput improvement and shipped as an installable VS Code extension

LatenC (2nd Place SFU CSSS Mountain Madness 2

Architected and implemented a production-ready game engine with Builder pattern replacing a 9-parameter constructor anti-pattern, reducing initialization errors and improving code maintainability through fluent interface design for flexible game composition with maps, entities, and rewards. – Engineered A* pathfinding algorithm from scratch using Manhattan distance heuristic and priority queue data structures achieving O(log n) average-case performance, implementing optimal pathfinding for enemy AI with obstacle avoidance across grid-based maps with 15 comprehensive edge case tests validating algorithm correctness. – Designed extensible architecture using Strategy pattern for pluggable pathfinding algorithms, enabling seamless integration of multiple AI behaviors through dependency injection while maintaining single responsibility principle across 35 unit tests – Implemented game loop system with tick-based processing handling player movement, enemy AI pathfinding, collision detection, and state management, coordinating interactions between 5+ module packages through clean separation of concerns with full documentation across 700+ lines of production code.