~/syed-mahi
Syed Mahi
title: Systems Builder Learning By Doing
location: Dhaka, Bangladesh
journey-start: Mid June 2025
university: Islamic University of Technology (IUT)
degree: BSc Industrial & Production Engineering
>|

$ cat about.md

# how I actually learn and build

about.md

I started my software journey in mid June 2025, and the first thing I built was IPE-24 Classroom. That project mattered because it was not a practice app. It had real users, daily pressure, and a real reason to exist.

My learning style is very hands-on. Reading syntax and loops in isolation feels boring to me, but building a real system, shipping it, watching it fail, and fixing it teaches me fast. Most of my progress has come from doing the work, not waiting to feel ready.

I want to become the best life problem solver I can by building efficient and scalable systems. Every time I notice a broken workflow or a new behavior pattern, a product idea starts forming. I am open to new learning challenges, and I am also interested in finding a strong seller or operator partner so the right systems can reach the world.

$ cat skills.json | jq '.techstack'

# the stack I actually use across projects

skills/languages

Languages

Java[███████░░░] 68%// Spring Boot services and backend workflows
TypeScript[████████░░] 82%// Frontend apps and typed product logic
Python[████████░░] 77%// Automation, APIs, ML experiments
SQL[████████░░] 79%// Schema design, querying, and relational modeling
Bash / Shell[████████░░] 76%// Automation, setup, and deployment scripting
skills/frameworks-product-stack

Frameworks & Product Stack

Next.js[█████████░] 87%// App Router, SEO, product landing pages
React[████████░░] 84%// Component systems, interactive product flows
Tailwind CSS[████████░░] 80%// Fast, product-focused styling systems
Spring Boot[██████░░░░] 64%// Workflow and marketplace backend work
skills/platforms-infrastructure

Platforms & Infrastructure

Vercel[████████░░] 83%// Fast frontend shipping and iteration
Docker[███████░░░] 74%// Portable local and deployment environments
Linux / VPS Deployment[███████░░░] 73%// Running and troubleshooting real systems
Git / GitHub[████████░░] 82%// Version control, iteration, documentation
skills/data,-auth-vector-layer

Data, Auth & Vector Layer

PostgreSQL[████████░░] 84%// Structured product data and workflow state
Prisma[████████░░] 78%// Schema-driven application modeling
Supabase[████████░░] 76%// Auth, Postgres, and fast product setup
Redis[███████░░░] 71%// Caching, queues, and acceleration layers
pgvector / Vector DB[███████░░░] 69%// Embeddings, retrieval, and AI knowledge search
Google OAuth[████████░░] 82%// Secure sign-in and delegated access flows
Gmail API[████████░░] 80%// Bulk sending and email workflow automation
skills/automation-ai-workflows

Automation & AI Workflows

Prompt-to-Structure Pipelines[████████░░] 78%// Turning messy input into usable system output
RAG / Embeddings[███████░░░] 71%// Knowledge retrieval and AI assistance
PptxGenJS[███████░░░] 74%// Programmatic deck generation
OpenCV / TensorFlow[███████░░░] 67%// Computer vision and prototype ML
Google AI Workflows[███████░░░] 70%// Applied model integration for product features

$ cat journey.log

# how the path has unfolded so far

journey/started-with-a-real-problem.log

Started With a Real Problem

Mid June 2025
@ IPE-24 Classroom

My journey started by building something my own batch could actually use every day, not by waiting to 'finish learning' first.

  • + First shipped project was IPE-24 Classroom
  • + Learned architecture, deployment, and product thinking under real pressure
  • + Real users forced me to care about reliability early
journey/learning-by-doing.log

Learning By Doing

2025 -- Present
@ Self-Directed Build Loop

My learning style is project-first. I understand ideas faster when they are attached to a real workflow, bug, user need, or business constraint.

  • + Syntax-first learning never held my attention for long
  • + I grow through shipping, breaking, debugging, and refining
  • + Every build becomes both a product and a lesson
journey/open-to-bigger-challenges.log

Open to Bigger Challenges

Now
@ Systems + Sales Partnership

I keep seeing new system ideas everywhere. I am especially open to a strong seller or operator partner so good systems can reach the people who need them.

  • + Interested in solving life problems with efficient systems
  • + Open to new learning challenges and hard product environments
  • + Looking for people who can help take strong builds into the market

$ tree ~/projects --dirsfirst

# projects grouped by current stage

projects/production.md

In Production

3 projects

Systems already solving real problems for active users and workflows.

~/projects/production/ipe24-classroom

IPE-24 Classroom

[PRODUCTION]

A daily-used class operating system for IUT students that centralizes announcements, class routines, resources, and AI-supported student workflows.

role: Founder, Product Designer, Full-Stack Builder
Next.jsPostgreSQLPrismaRedisTelegram BotFastAPI
  • -> My first project and the place where my journey started in mid June 2025
  • -> Used daily by 60+ students as a shared academic source of truth
  • -> Combines web, automation, and AI support instead of being just a portal
  • -> Built around reliability, fast publishing, and guardrails for real student use
case study live
~/projects/production/email-automator

Email Automator

[PRODUCTION]

A Gmail-based bulk email tool for sending personalized announcements, grades, and reports with preview-first safety and structured data merging.

role: Product Builder
Next.jsNextAuthGoogle OAuthGmail APIHandlebars
  • -> Designed for high-stakes sending where one bad batch breaks trust
  • -> Merges messy CSV data into deterministic recipient workflows
  • -> Uses the sender's own Google identity instead of acting like a separate mail provider
  • -> Focused on safety, preview control, and low-friction operations
case study live
~/projects/production/slidecrafter-ai

SlideCrafter AI

[PRODUCTION]

A local-first AI presentation compiler that turns raw context into editable decks by generating structured slide instructions instead of vague slide text.

role: Product Builder
ReactTypeScriptVitePptxGenJSGemini
  • -> Treats deck generation as a compiler problem, not just prompt decoration
  • -> Validates layout structure before export to avoid broken slides
  • -> Supports research-heavy workflows with file ingestion and repair loops
  • -> Keeps final deck compilation close to the user for faster iteration
case study live
projects/inventory.md

Inventory

2 projects

Shipped products and operational builds that are ready to be positioned, sold, or extended.

~/projects/inventory/the-matrix-hours

The Matrix Hours

[INVENTORY]

A premium watch storefront shaped around trust, curated inventory, and a low-friction buying flow that fits the business better than overbuilt commerce complexity.

role: Product Builder
Next.jsReactGoogle SheetsWhatsApp Checkout
  • -> Built a product catalog around business reality instead of forcing heavyweight ecommerce
  • -> Uses Google Sheets as an owner-friendly inventory source with a normalization layer
  • -> Balances luxury presentation with operational simplicity
  • -> Designed for staged certainty rather than unnecessary infrastructure early
case study live
~/projects/inventory/tutors-hub

Tutors Hub

[INVENTORY]

A tutoring marketplace that behaves like an operations system, with workflow-heavy matching, verification, and human recovery paths built into the product.

role: Product Builder
Next.jsSpring BootPostgreSQLSupabaseFirebase
  • -> Designed as a workflow platform, not a simple tutor listing site
  • -> Models matching, routing, verification, and payout logic as business state transitions
  • -> Separates onboarding UX from security-sensitive backend workflows
  • -> Built around trust boundaries and operational recovery paths
case study live
projects/development.md

Development

2 projects

Projects I am actively sharpening, validating, or growing into stronger products.

~/projects/development/krishi-dristi

Krishi Dristi

[DEVELOPMENT]

An offline-first agricultural intelligence prototype that maps fragmented farm plots, detects crop issues, and works under real field constraints in Bangladesh.

role: Prototype Builder
PythonOpenCVTensorFlowMobileNetV2
  • -> Built around smallholder realities instead of imported precision-ag assumptions
  • -> Offline-first because the target environment cannot depend on stable internet
  • -> Combines classical vision and ML where each actually fits
  • -> One of my first deeper system-design plus ML experiments
case study
~/projects/development/logbook

LogBook

[DEVELOPMENT]

An academic SaaS direction focused on policy-driven grading, offline-first educator workflows, and a stronger foundation for trustworthy record keeping.

role: System Designer
Next.jsTypeScriptPostgreSQLSupabase
  • -> Treats grading policy as data instead of hardcoded assumptions
  • -> Explores offline-first academic workflows beyond a simple attendance app
  • -> Shaped by real educator pain instead of generic edtech clichés
  • -> Focused on getting the long-term architecture right early
case study
projects/pre-processing.md

Pre-Processing

3 projects

Ideas in workflow discovery, system framing, and early product architecture.

~/projects/pre-processing/ops

OPS

[PRE-PROCESS]

A pre-processing concept for an ad agency management system covering client intake, approvals, delivery flow, and internal visibility.

role: Problem Framing & System Design
Next.jsTypeScriptPostgreSQL
  • -> Currently in workflow discovery and system framing
  • -> Focused on making agency execution more structured and traceable
  • -> Exploring how sales, delivery, and reporting should connect cleanly
~/projects/pre-processing/blird

Blird

[PRE-PROCESS]

A hyper-local blood donation helper concept designed to reduce the time between a request, the right donor match, and actual life-saving coordination.

role: Problem Framing & System Design
Next.jsSupabaseMapsNotifications
  • -> Centered on urgent local coordination instead of a broad social feed
  • -> Exploring trust, donor availability, and geographic relevance as core constraints
  • -> Shaping the system around fast human response when time matters most
~/projects/pre-processing/iut-classroom-booking

Classroom Booking System for IUT

[PRE-PROCESS]

A pre-processing systems idea for making classroom availability, booking requests, and approval flows cleaner for IUT.

role: Problem Framing & System Design
Next.jsPostgreSQLCalendar WorkflowsRole-Based Access
  • -> Aims to reduce manual friction in institutional room scheduling
  • -> Exploring booking conflicts, approval paths, and operational visibility
  • -> Still in the system-mapping stage before implementation

$ find './Case study' -name '*.md' | sort

# deep dives built from the actual case study folder

~/case-studies

Write-ups grounded in the real project notes, architecture thinking, and product lessons behind the systems above.

$cd ./case-studies && ls -la_

$ cat contact.sh

# open to building, learning, and partnerships

contact.sh

# If you want to build useful systems together, or you are a strong seller/operator who sees the same opportunity, reach out.

echo "hello" | mail syedmahi@iut-dhaka.eduopen https://github.com/syedmahi-devopen https://www.linkedin.com/in/syed-mahmudur-rahman-mahi-080242207/
# Response time: usually within 24 hours