We bridge the gap between intelligent software and real-world systems — from LLM pipelines and RL agents to hardware integration and 3D spatial computing. originull is what happens when those two worlds collide.
Builds AI-native systems that turn experimental models into production-grade software. From LLM automation pipelines and RAG systems to multi-modal agents — he designs infrastructure where learning, reasoning, and execution operate as one layer.
Founder of Dappi, leading end-to-end product execution and translating complex AI capabilities into scalable applications used by businesses.
Designs intelligent control systems that connect decision-making algorithms to real-world execution — from training RL agents for combat strategies, to proximity-based authentication, to translating software commands into deterministic hardware behavior.
Interested in scalable architectures, adaptive autonomy, and turning experimental AI into deployable infrastructure.
Production LLM extraction pipeline with validation layers for enterprise invoice processing.
Scalable Retrieval-Augmented Generation systems with deployment guardrails and multilingual fine-tuning.
Automated grading system using NLP and deep learning for structured assessment pipelines.
AI-driven anomaly detection system for ESG and security monitoring using XGBoost.
Graph-based semantic matching algorithm for intelligent job–candidate pairing.
End-to-end mobile platform with payments, operations, and backend architecture.
Reinforcement-learning agent trained to discover efficient combat strategies through human-like input channels.
System that translates software commands into deterministic hardware behavior across consumer devices.
Programmatic control of Android devices using APIs and debugger tools for automation and testing.
Real-time gesture recognition to control audio equalization system-wide across the operating system.
Graph-based semantic matching algorithm for intelligent job–candidate pairing.
Creative 3D modeling and design projects exploring spatial computing and visual media.