Zaiff

Zaiff Khan

Software Engineer

Backend Systems • Full-Stack Applications • AI Automation

Resume (PDF)

Full-Stack Imageboard Platform designed for high-concurrency asset processing and automated content ingestion.

FastAPI Python SQLite (WAL) Jinja2 Asynchronous I/O REST API
  • Data Pipeline & Automation: Developed a custom asynchronous task worker system to automate the ingestion, tagging, and processing of ~200,000 high-resolution media assets with zero database locking.
  • API Development: Designed and deployed a secure, key-authenticated REST API utilizing FastAPI dependencies, implementing custom in-memory rate-limiting protocols and asynchronous audit logging.
  • System Architecture: Architected a high-performance SSR web application engineering a multi-database SQLite infrastructure (WAL mode) optimized for high-concurrency read operations.
  • Content Routing & SEO: Engineered a dynamic subdomain system with automated strict-SFW filtering, utilizing canonical tags and JSON-LD schemas to preserve root-domain SEO authority.

Autonomous Multimodal AI Agent (Nova)

An intelligent autonomous agent featuring multi-tier memory systems, self-modifying protocols, and hardware/API integration for environment interaction.

Python Vector DB Vision APIs Discord API Arduino Asynchronous Systems
  • Tiered Memory Architecture: Architected a tiered memory system balancing standard RAM with vector-based Medium-Term/Long-Term Memory (MTM/LTM) alongside a self-modifying "Addendum" protocol, allowing the agent to persist context and autonomously evolve its core instructions over time.
  • Cognitive Reasoning Loops: Engineered a multi-stage "Thought Loop" reasoning chain that utilizes weighted personality sliders and an algorithmic three-tier voting system to determine when to actively initiate contact versus remaining silent.
  • Multimodal Tool-Use: Integrated tool-use across Vision APIs, Discord voice/image protocols, and physical Arduino hardware, enabling the agent to "spectate" virtual screens and programmatically interact with its physical environment.

Mage TD

Completed Project

A fully playable, 6-level 2D Tower Defense game engineered in Godot, featuring custom upgrades, unique enemy behaviors, and hand-drawn sprite animations.

Godot Engine GDScript Object-Oriented Programming State Machines Performance Optimization
  • Game Architecture & Systems: Programmed a robust gameplay loop across 6 distinct levels, implementing modular custom tower structures, persistent in-game/meta-progression upgrade systems, and dynamic economy scaling.
  • Behavioral State Machines: Engineered unique enemy AI behaviors and wave spawning logic using clean object-oriented architecture in GDScript, ensuring decoupled, maintainable code across varying entity types.
  • Performance & Asset Pipeline: Managed a complete creative and technical asset pipeline, designing handmade sprite art and multi-frame animations while optimizing resource management to ensure the engine runs flawlessly on low-spec hardware.