I'm Tamiz. I build systems that scale.
Full-stack engineer with 2+ years building production systems — from AI-powered tools to real-time platforms and e-governance infrastructure. Currently engineering at Star Connect.
I build production systems that solve real problems.
I'm Tamiz Uddin, a full-stack engineer with 2+ years of experience shipping software that's correct first, fast second, and maintainable always. I work across the entire stack — designing APIs, modeling databases, and crafting interfaces that feel effortless.
My recent work spans AI-powered developer tools, real-time platforms, and e-governance infrastructure. I care deeply about clean architecture and the trade-offs behind every technical decision, and I enjoy the space where frontend polish meets backend resilience.
When I'm not building, I'm exploring new tools in the AI and web ecosystem — and turning the best ideas into real products.
Location
Remote · Bangladesh (UTC+6)
Current Role
Software Engineer @ Star Connect
Focus
Full-stack · AI Integration
Availability
Open to freelance & contract
Engineering Stack
The languages, frameworks, and infrastructure I use daily to build production systems.
Languages
Frontend Architecture
Backend & Systems
AI & Innovation
Tools & Infrastructure
Rust & WebAssembly
Exploring low-level performance for compute-intensive frontend tasks...
Expertise & Architecture
How I approach each layer of the stack — from interactive frontends to resilient backend systems.
Frontend Architecture
- React & Next.js EcosystemApp Router patterns, SSR, ISR, and edge caching strategies for optimal TTFB.
- State & Data FlowComplex state machines, optimistic UI updates, and intelligent data fetching.
- Design SystemsBuilding scalable, accessible UI libraries with Tailwind CSS and Framer Motion.
Backend & AI Integration
- GenAI OrchestrationIntegrating Google Gemini Pro for codebase analysis, trend synthesis, and automated insights.
- Security & AuthJWT implementations, OAuth flows, and strict RBAC authorization logic.
- Database ModelingPostgreSQL/MongoDB schema design, complex query optimization, and indexing.
DevOps & Infra
- CI/CD PipelinesAutomated testing, linting, and zero-downtime deployment workflows.
- Cloud InfrastructureAWS/Vps provisioning, container orchestration, and serverless functions.
- Caching & PerformanceRedis implementation, CDN configuration, and intelligent database caching.
Featured Systems & Architecture
Property Tax Management System
Municipalities and local governments often struggle with manual property tracking, lack of spatial awareness for land boundaries, inefficient billing processes, and disconnected payment systems, which lead to low tax collection rates and poor transparency.
A modern Client-Server architecture utilizing a React SPA built with Vite and TypeScript on the frontend, and a Node.js/Express REST API on the backend. The system stores spatial and structured data in MongoDB, uses PM2 for backend process management, and leverages cron jobs for background processing.
VelocityPro - Internet Speed Test
Users need an accurate, modern way to test their internet speed and understand connection quality. Existing speed test tools are often outdated, lack real-time visualization, or don't provide meaningful context for different use cases like gaming, video calls, and 4K streaming.
Client-server Next.js application with Firebase Firestore backend. The frontend (React 19, TypeScript) handles real-time speed testing via API endpoints for download/upload/ping measurements. Cloudflare's edge network provides global test server distribution. Firebase Firestore manages shared leaderboard data and test history. API routes proxy connection info and provide self-hosted speed test endpoints.
Multilingual News Platform & CMS
Global news platforms require high-performance, SEO-optimized delivery across languages while providing editorial teams with powerful rich-media management tools.
Next.js application utilizing the App Router and a custom Node.js server. Backed by MongoDB (Mongoose) and Cloudinary for CDN-based media delivery.
Live TV - Real-time Streaming Platform
Traditional TV streaming requires complex infrastructure to handle concurrent viewers, protocol compatibility, and real-time viewer state synchronization. Existing solutions are either closed-source or lack modern web capabilities.
Next.js full-stack application with a three-layer architecture: Frontend (React 19 with Tailwind CSS for UI), Backend (Next.js API routes with Mongoose/MongoDB), and Real-time Layer (Socket.io with Redis adapter for scalable WebSocket connections). HLS.js and MPEGTS.js handle different streaming protocols client-side.
How I Build Systems
Every project follows the same loop: understand the problem deeply, design for the edge cases, build incrementally, and ship with confidence.
01Understand the Problem
I start by mapping business logic, user flows, and edge cases — understanding what the system actually needs before writing any code.
02Design the Architecture
I draft the system design, choose the right database schema, and define API contracts and domain boundaries.
03Build Incrementally
I build features in decoupled modules, focusing on typed interfaces, clean code, and composability.
04Test & Harden
I load-test critical paths, audit dependencies, and make sure authentication and authorization are airtight.
05Ship & Monitor
I set up CI/CD for zero-downtime deployments and connect monitoring for proactive issue resolution.
AI & Automation
I integrate LLMs and machine learning into core product logic — not as wrappers, but as first-class features.
Gemini AI Orchestration
Utilizing Google's Gemini Pro API to build intelligent features like automated codebase auditing (RepoInsight) and real-time trend synthesis (TrendPulse).
Intelligent Productivity
Designing focus-driven applications like the AI-Enhanced Pomodoro, which leverages LLMs to break down complex goals into actionable, bite-sized tasks.
Predictive Analytics
Deploying machine learning workflows to analyze cross-platform trends and visualize insights through interactive, time-series dashboards.
Impact & Testimonials
Numbers from the work I've shipped, and words from the people I've built with.
“Tamiz is a rare talent who deeply understands both the structural integrity of a system and the nuances of user experience. He doesn't just write code; he engineers solutions.”
“The level of technical depth and attention to detail Tamiz brings to every project is exceptional. He consistently delivers high-performance architectures.”
Engineering Notes
From Code to Context: How OpenViking's Self-Evolving Memory and 'rea's Reverse Engineering are Redefining the AI Agent Lifecycle
Exploring how OpenViking's self-evolving memory architecture and 'rea's reverse engineering capabilities transform AI agents from static scripts into dynamic, context-aware systems.
Beyond the Prompt: Building Safe, Local Execution Environments for Agentic Tooling with Rust and smolvm
Learn to build secure, sandboxed execution environments for AI agents using Rust and smolvm — covering capability-based security, WASM isolation, and production hardening.
The Post-Lock-in Agent Era: How 'gentle-ai' and 'agentic-awesome-skills' Are Reclaiming Control from Proprietary Coding Agents
Analyze the emerging open-standard movement. Learn how 'gentle-ai' and 'agentic-awesome-skills' provide portable, local-first agent architectures that break proprietary lock-in.
Let's build something meaningful.
I'm currently open to freelance projects, SaaS collaborations, and contract engineering roles. If you have a system to build or a team that needs an extra engineer — let's talk.