Danial Kordmodanlou

Machine Learning Associate (MLA) at Vector Institute

Partnered with DXTR to build a real-time VR firefighter training system detecting skill deviations via headset telemetry. Architected a Deviation Engine processing sliding windows of 3D skeletal telemetry, computing MPJPE and quaternion angular distances to assess deviations aligned to training standards. Developed an LLM-driven Coaching Layer utilizing RAG (FAISS similarity search) and Ollama/LangChain to generate natural language feedback grounded in manual excerpts. Engineered a full-body pose collection pipeline, mapping headset telemetry to NTU-25 skeletons and normalizing frames for cross-platform robustness. Collaborated on data integration and structured evaluation of generative AI pipelines for high-fidelity training scenarios.

Core Contributor – Unity/C# Developer at DreamForge

Core contributor to an AI-driven game engine that procedurally generates 3D game worlds from text prompts. Implemented a procedural wall cutout system using CSG boolean subtraction with four silhouette strategies (convex hull, radial ray casting, voxel tracing, alpha-shape), connected-component filtering, and ear-clipping triangulation. Built CI/CD pipelines using Docker, Azure Container Instances, .NET, GitHub Actions, and Slack webhooks for automated exception analysis, PR creation, and deployment notifications. Architected a multi-mode camera system using Cinemachine with adaptive framing, multi-candidate occlusion-aware selection, bone-based NPC framing, spherecast fading, and room-geometry yaw constraints. Designed a polygon-based room shape Generator with grid-based flood fill validation.

Research Assistant & Teaching Assistant at BioMotion Lab, York University

Conducting VR and motion capture experiments using Meta Quest, ARKit, Unity3D, and Unreal Engine, exploring depth perception by isolating stereopsis and motion parallax. Presented research on illusory parallax in VR at ECVP 2025. Teaching Object-Oriented Programming (Java) for 5 semesters, emphasizing system design patterns and optimization.

Software Developer at TectoTrack

Built simulation frameworks for high-traffic environments, integrating fault-tolerant pathfinding for adaptive crowd movement. Developed real-time logging and monitoring dashboards to track system health and agent behavior metrics.

Lead Unity3D Game Developer at Techu

Led the development of 'Techu,' a cross-platform 2D card game using DQN+Monte Carlo Tree Search for AI agents, and PlayFab Azure for matchmaking and data persistence. Integrated Firebase monitoring for real-time system diagnostics and automated crash alerts.

Volunteer Full Stack Developer at IAESTE

Maintained and enhanced IAESTE’s web platform using Vue.js and Django, improving performance and user experience.

Technical Staff Member & Teaching Assistant at Amirkabir University of Technology (AUT)

Technical staff for AUT Game Development Events and TA for AI, Advanced Java, C Programming, and Operating Systems.