Gaming AI Agent

Python, PyTorch, NumPy, Mathplotlib

Engineered a real-time data acquisition system capturing synchronized 30 FPS RGB game frames and keyboard/mouse actions from a live TF2 environment, generating a large-scale imitation learning dataset for supervised policy training.

Built a high-throughput preprocessing and tensorization pipeline optimized for GPU training, converting raw visual streams into structured model-ready inputs with parallel worker execution.

Reduced end-to-end preprocessing latency from ~120 ms to ~40 ms per frame through multiprocessing, batching strategies, and pipeline optimization, increasing overall training throughput by ~2.5x.

Trained a supervised neural network on an RTX 4060 over multiple days of continuous training, learning a direct mapping from visual state representations to player action distributions.

Scaled model input representation using temporal frame stacking (2.5x context window) and expanded action-space encoding, improving temporal stability and reducing prediction noise during gameplay imitation.

Autonomous Vehicle Autopilot

Lua

Engineered a modular autopilot control system implementing real-time navigation logic for ascent, cruise, and landing phases in a simulated environment.

Designed sensor-driven feedback loops for obstacle avoidance and trajectory correction, enabling adaptive real-time path planning.

Increased control-loop execution frequency from 15 Hz to 60 Hz by optimizing update logic and reducing per-cycle computational overhead.

Built reusable control modules for throttle, steering, and waypoint navigation, enabling scalable reuse across vehicle configurations.

Custom Transformer Language Model

Python, PyTorch

Designed and implemented a 24-layer transformer language model with 16 attention heads and 1024-dimensional embeddings for large-scale sequence modeling and text generation.

Engineered a full training pipeline on a 15GB text dataset using an RTX 4060 GPU, achieving stable convergence over extended multi-day training runs.

Built a custom tokenizer supporting up to 50,000-token vocabulary while preserving semantic structure and improving contextual representation quality.

Implemented optimized training stack using AdamW optimization, cross-entropy loss, and sequence truncation strategies to stabilize long-context learning behavior.

Developed inference engine with temperature scaling and top-k sampling, producing coherent and contextually consistent text outputs across varied prompts.

Spotify API Based Multiplayer Web Game

HTML, CSS, JavaScript

Engineered a real-time multiplayer game system integrating Spotify’s REST API with a WebSocket-based backend for low-latency state synchronization across concurrent clients.

Designed a custom event-driven networking architecture supporting up to 25 simultaneous users, hosted on my home server.

Implemented event batching and state-diff synchronization to reduce client-side render jitter by ~25% under concurrent load.

Built a responsive frontend using adaptive layout logic to ensure consistent rendering performance across devices and screen sizes.

Optimized client update loop and DOM interaction patterns to minimize unnecessary re-renders and maintain stable real-time UI behavior.

Secure Banking CLI

Python

Designed a secure CLI-based banking system with separation of authentication, transaction processing, and persistence layers.

Implemented salted SHA-256 password hashing with login throttling and verification safeguards to strengthen account security.

Designed a precision-safe financial transaction engine using Python Decimal, eliminating floating-point rounding errors in monetary operations.

Built JSON-based persistent storage with structured transaction logging and ISO-8601 timestamps for auditability and recovery.

Articulating Robotic Arm

Python

Engineered a 3-DOF robotic arm with MG996R servos and a PCA9685 PWM driver, using embedded systems programming on a Raspberry Pi Pico.

Increased servo response consistency by ~15% using optimized PWM timing.

Designed and 3D-modeled all mechanical components, applying precision tolerance control and iterative prototyping for smooth servo actuation.

Integrated manual and automated command modes, supporting scripted movements and real-time control input for flexible operational scenarios.

KSP KerboScript Boostback Landing

kOS, KerboScript

Engineered a comprehensive Kerboscript-based autopilot system for Kerbal Space Program that automates flight phases from ascent to landing.

Designed a telemetry-driven ascent algorithm using control-system logic to dynamically adjust throttle and pitch, improving trajectory accuracy and efficiency.

Implemented a boostback routine using trajectory-prediction algorithms to synchronize thrust and steering for accurate return-to-launch-site maneuvers.

Developed an entry burn sequence paired with an aerodynamic guidance system, employing real-time pitch and yaw error corrections to align the vehicle with predetermined landing coordinates during atmospheric re-entry.

Engineered a suicide burn landing control-loop algorithm that computes stopping distances on-the-fly, modulating throttle to ensure a controlled landing.

Improved landing accuracy to within ±0.25 m of target coordinates.

Simple Cypher

C++

Developed a C++ command-line encryption tool using fundamental cryptography techniques, including Caesar shift and XOR byte-wise transforms.

Implemented efficient character-rotation logic to handle configurable Caesar cipher shifts with consistent output integrity.

Built a reversible XOR routine with an 8-bit key, using optimized bitwise operations and hex output for non-printable character support.

Reduced runtime by ~45% by simplifying character-rotation logic.