Available for opportunities
⌖ Dhaka, Bangladesh · Remote-friendly AI Workflows · SaaS · Marketplaces

Kazi Habib

I turn risky AI feature ideas into tested production slices — helping startups and product teams move from an uncertain workflow to defined behavior, full-stack implementation, evaluation, guardrails and measurable value.

  • 9+ Years Experience
  • 52K+ Udemy Learners
  • 2 Live Products
  • 30 AI Evaluation Cases
Kazi Habib, AI Product Engineer
Kazi Habib AI Product Engineer
Open
Next.js·React·TypeScript·Node.js·PostgreSQL·MongoDB·OpenAI API·Structured outputs·RAG·Evaluations·Guardrails·Firebase·Kotlin·Docker·Vercel

About Me

Product judgment and engineering ownership in one person.

Engineering robust, evidence-driven solutions across the full stack.

I am Kazi Habib, an AI Product Engineer and full-stack product builder based in Dhaka, Bangladesh. For more than nine years I have built, taught and operated software across web, backend, mobile and real-time systems.

I have launched independent products including Bhara.com and ParenTek, delivered workflow and backend systems for international teams, and taught programming to more than 52,000 Udemy learners. That mix shaped how I work: make the problem explicit, explain decisions clearly, build the narrowest useful system and measure what happens after it ships.

My current work focuses on ProductOS and SupportSlice — practical demonstrations of evidence-gated product decisions and reliable, human-reviewed AI workflows.

52K+ Learners taught
8 Published courses
2 Live products operated
3 Regions delivered across

Skills

Technical Arsenal.

Tools and technologies I use to define, build, evaluate and ship reliable AI products.

Frontend

  • React
  • Next.js
  • TypeScript
  • JavaScript

Backend

  • Node.js
  • Express.js
  • REST APIs
  • WebSockets

Databases

  • PostgreSQL
  • MongoDB
  • SQL
  • Drizzle ORM

Applied AI

  • OpenAI API
  • Structured outputs
  • RAG
  • Guardrails
  • Evaluations
  • Permission-aware retrieval

Mobile & Platform

  • Kotlin
  • Native Android
  • Firebase
  • FCM

Cloud & DevOps

  • Docker
  • Git & GitHub
  • Vercel
  • AWS

Experience

My professional journey building software and teaching it at scale.

A decade of product delivery, international client work and technical education — always ending in something deployed and measured.

KH

Independent Senior Software Engineer

Product delivery & product ownership · 9+ years

9+ years
  • Take products from first commit through deployment — MVPs, AI-generated codebases, authentication, APIs, payments and third-party integrations.
  • Own architecture, implementation, debugging and delivery while working directly with founders and distributed teams.
  • Launched and operated live products end to end, including the Bhara.com marketplace and the ParenTek family-safety platform.
ReactNext.jsNode.jsTypeScriptPostgreSQLMongoDBDockerVercel
B

Software Engineer — Banff Advisors

Canada · Workflow automation

  • Replaced manual internal business processes with automated pipelines and production workflows.
Node.jsAutomationWorkflows
HM

Backend Engineer — heartattack.me

Middle East · Multi-country restaurant platform

  • Developed backend systems supporting restaurant ordering, menus and operations across three countries.
Node.jsExpress.jsMongoDBREST APIs
U

Technical Instructor — Udemy

52,000+ learners · Eight published courses

Profile ↗
  • Teach practical software development — full-stack engineering, WebRTC, real-time communication, React Native, AWS and AI-assisted SaaS development.
  • Translate complex technical ideas into clear, project-based guidance for more than 52,000 learners.
TeachingWebRTCReact NativeAWSAI-assisted development

Projects

Featured projects — proof through working systems.

Product judgment, AI behavior and full-stack delivery, shown in products you can inspect.

ProductOS evidence-gated product decision workspace 01 LIVE APP

EVIDENCE-GATED PRODUCT ENGINEERING

ProductOS

Flagship project · In active development

A decision workspace that moves one product or feature from uncertainty to a defensible Scale, Change or Stop decision — with evidence, behavior and approval gates kept visible in one durable workflow.

  • Evidence provenance and assumption ledger
  • Approval gates and durable decision artifacts
  • AI/manual responsibility boundaries
  • Verification and value-measurement paths
Next.jsTypeScriptPostgreSQLDrizzle ORMStructured outputsEvaluations
SupportSlice interface showing grounded drafts, human decisions and citation checks 02 LIVE DEMO

GROUNDED CUSTOMER-SUPPORT COPILOT

SupportSlice

Portfolio prototype · Synthetic data only

Drafts B2B SaaS support replies from approved evidence while keeping the support agent accountable for every final decision.

  • Permission-aware retrieval
  • Exact citation validation
  • Abstention and safe escalation
  • Prompt-injection defenses
  • 30 synthetic evaluation cases
  • Human approval and audit trail
Next.jsTypeScriptRAGOpenAI APIEvaluations

Transparent by design: the company, tickets and metrics are synthetic. Demonstration behavior is measured; business impact remains a hypothesis for a real pilot.

Bhara.com broker-free rental marketplace homepage 03 LIVE PRODUCT

BROKER-FREE RENTAL MARKETPLACE

Bhara.com

Live product · 200+ listings

A broker-free rental marketplace for Bangladesh, built and launched end to end and operated as a live product.

  • Listings, search and interactive maps
  • Authentication and owner onboarding
  • APIs and production deployment
  • 200+ property listings onboarded
ReactNode.jsMongoDBMapsAuth
ParenTek family-safety dashboard 04 LIVE PRODUCT

BILINGUAL FAMILY-SAFETY PLATFORM

ParenTek

Live · Native Android + Next.js web

Device analytics, location visibility, app management and alerts for families — with Firebase workflows and AI-assisted insights that keep human judgment in risk-sensitive decisions.

  • Native Kotlin Android application
  • Next.js parent dashboard
  • Location visibility and app management
  • AI-assisted insights with human review
KotlinAndroidFirebaseNext.jsFCM

Community & Teaching

A glimpse into what I ship and teach.

TECHNICAL INSTRUCTOR

52,000+

Udemy learners · 8 published courses

Practical, project-based courses across full-stack engineering, WebRTC, React Native, AWS and AI-assisted SaaS development.

View instructor profile ↗

OPEN SOURCE

kmhabib71

GitHub · ProductOS and SupportSlice source

Full source for the evidence-gated decision workspace and the grounded support copilot — tests, guardrails and evaluation sets included.

View GitHub ↗

LIVE PRODUCTS

Bhara · ParenTek

Two products I founded and operate

A live rental marketplace with 200+ listings and a bilingual family-safety platform spanning Android and web.

Bhara.com ↗ ParenTek ↗

How I Build Systems

Core principles that guide my engineering decisions.

01

Evidence over enthusiasm

Facts, assumptions, unknowns and counter-evidence stay separated. Scale, Change or Stop decisions are made with proof, not demo excitement.

02

Narrowest useful slice

Build the smallest complete end-to-end workflow that can show value — never an oversized platform before the problem is real.

03

Behavior defined first

Specify what the system does, must not do and when a human decides — before writing the integration code.

04

Failure tested early

Evaluation sets cover normal cases, edge cases, unsafe actions and provider failures — so problems surface before production.

05

Humans stay in control

Role-based access, approval gates and audit trails keep consequential decisions with accountable people.

06

Value measured after ship

Deployment is not the finish line. Pilot analytics and honest iteration decide whether the next investment is justified.

Key Achievements

Milestones and impact delivered.

★

Technical education at scale

52,000+ learners across eight published Udemy courses — turning complex engineering topics into clear, project-based guidance.

★

Live products from first commit

Founded, built and operated Bhara.com with 200+ property listings and ParenTek across native Android and the web.

★

Evidence-gated AI workflows

ProductOS decision workspace with explicit approval gates, plus a SupportSlice evaluation suite of 30 synthetic cases across eight dimensions.

Get in Touch

Let's turn an uncertain idea into a reliable product.

Open to remote contract work and full-time AI Product Engineer roles. If you are building an AI workflow and need someone to define its behavior, build the narrowest useful slice, evaluate its failures and measure the value — let's talk.

Kazi Habib AI Product Engineer
Available
Email
GitHub
github.com/kmhabib71
LinkedIn
—
Udemy
udemy.com/user/KAZI-habib

Send a message

Describe the workflow, the outcome you want and the most serious failure you need to avoid.

Submitting opens your email app addressed to Kazi Habib — no message is stored on this site.