Global State Management R&D | LLM Systems | Real-Time Audio Engineer

I’m a research-driven full-stack engineer focused on designing new paradigms for global state management, building LLM-powered systems, and engineering low-latency real-time audio pipelines.
I work at the intersection of frontend architecture, AI systems, and performance-critical software.

• Designing abstractions beyond Redux
• Configuration-driven state architecture
• Function storage and controlled execution inside global state
• Deterministic updates, garbage collection, and IFTTT-style triggers

• Conversational AI with real-time feedback loops
• Streaming token pipelines
• Context management and evaluation logic
• AI interviewers and reasoning assessment systems

• Real-time STT and TTS streaming
• Whisper and Azure Speech integration
• Noise reduction using RNNoise
• Low-latency audio windowing and buffering strategies

A research-driven global state management system designed to reduce Redux complexity while preserving power and scalability.

Key ideas:

• Zero-boilerplate global state initialization
• Instance-based and lightweight state usage modes
• Configurable behavior instead of rigid conventions
• Support for storing and managing functions in global state
• Deterministic state updates with optional side-effect triggers

Repository: https://github.com/shadmax22/react-usethis

An LLM-powered interview system designed to simulate real technical interviews.

Features:

• Conversational interview flow
• Dynamic follow-up questions based on candidate responses
• Coding playground with live evaluation
• Critical thinking and reasoning assessment
• Scoring and feedback generation

Repository: [PRIVATE]

A low-latency audio pipeline for human–AI interaction.

Focus areas:

• Streaming microphone input
• Real-time Whisper STT
• Azure-based TTS with sentence-level streaming
• Noise suppression using RNNoise
• Audio windowing for fast transcription turnaround

Repository: [PRIVATE]

• React.js, Next.js, Vite
• TypeScript, JavaScript

• Node.js
• PHP
• Python
• System design and API architecture

• LLMs
• Whisper
• Azure Speech (STT / TTS)
• RNNoise
• Streaming and buffering pipelines

• Git, GitHub
• VPS, Linux
• Performance optimization and low-latency systems

• Started professional software development right after class 12
• Currently working as a Lead Full-Stack Developer
• Actively researching state systems, AI behavior, and human–AI interfaces
• Strong focus on reducing abstraction cost while increasing system power

• BSc in Data Science and AI
• Continuous experimentation with system internals and architecture trade-offs

• GitHub: @shadmax22
• Portfolio / Blog: mdshad.com

I research, design, and build systems that sit beneath the UI, where performance, determinism, and intelligence actually matter.