Research-backed · Beginner-friendly · August 2026

Learn the things a professional Software Engineer actually needs — in the right order.

এই guide তোমাকে framework list মুখস্থ করাবে না. It explains what each skill is, why you need it, how deep you need to learn it, and when to move to the next step. Goal: requirement বুঝা থেকে production system design, build, test, secure, deploy, debug and improve করা.

FoundationWeb · JS/TS · Git · Debugging

BuildReact · Node · APIs · PostgreSQL

EngineerTesting · Security · Production · Architecture

AI EraAI speed + human verification

00 · How to use

এই guide কীভাবে follow করবে

One simple rule: Learn in order. আগের core step weak রেখে “advanced” topic-এ jump করবে না.

MUST DEEP

Strongly learn

Explain + build + debug + test করতে হবে. AI ছাড়া basic reasoning পারতে হবে.

MUST PRACTICAL

Real work level

Real project-এ use করতে পারবে. Internal theory expert হওয়া লাগবে না.

WORKING

Understand + use docs

Concept clear থাকবে; docs দেখে safe implementation করতে পারলেই enough.

BASIC ONLY

Know what & why

এটা কী/কেন/কখন লাগে জানলেই এখন হবে. Deep study পরে.

Completion rule: Video শেষ = topic শেষ না. Topic শেষ when you can explain → implement → debug → verify.

Beginner practice loop

প্রতিটা topic একইভাবে finish করবে. First understand the idea, then build something tiny, break it on purpose, debug it, test it, and explain the trade-off in your own words.

Learn the concept
→ Build a small version
→ Add one realistic edge case
→ Debug one failure
→ Write/record a short explanation
→ Add it to your notes or project README

Beginner warning: শুধু course দেখা বা copy-paste project শেষ করা skill proof না. Proof হলো তুমি নিজে requirement বুঝে small change safely করতে পারো.

01 · Big picture

Professional Software Engineer আসলে কী করে?

Software Engineer শুধু code writer না. IEEE SWEBOK-এর current software-engineering body of knowledge requirements, design, construction, testing, operations, maintenance, quality, security, configuration management এবং professional practice-কে discipline-এর অংশ হিসেবে রাখে.

UnderstandProblem

DesignSolution

BuildCode

TestCorrectness

SecureRisk

DeployProduction

ObserveReality

ImproveMaintain

Your target: “I know React” নয়. Target হলো: “I can take ownership of a software feature/system from requirement to production and explain my technical decisions.”

02 · Complete learning path

কোনটার পরে কোনটা শিখবে — and why

Step

Learn

Why this comes now

Time

1

Requirements & SDLC

Wrong problem solve না করে software work-এর full lifecycle বুঝবে.

2–3 days

2

HTML/CSS + Web Basics

Semantic UI, responsive layout and browser/network fundamentals বুঝবে.

2 weeks

3

JavaScript → TypeScript

Your main programming foundation and type-safe large codebase skills.

3 weeks

4

Git & Debugging

Team code safely change and evidence দিয়ে bug fix.

1 week

5

Frontend

Correct, accessible and maintainable React/Next user experience.

3 weeks

6

Backend & APIs

Business rules, integrations, validation, auth.

4 weeks

7

Database

Correct data model, transactions, indexes and reporting.

3 weeks

8

Design + Quality

Maintainable, tested, secure software.

4 weeks

9

Production

Docker, CI/CD, cloud, monitoring, recovery.

4 weeks

10

System Design + Architecture

Now you have real problems to reason about at system level.

3 weeks

11

AI-Era Engineering

AI makes you faster without removing your engineering control.

2 weeks

12

Professional Practice

Communication, reviews, docs, estimation and ownership.

Continuous

03 · Current market

2026-এ companies repeatedভাবে কী চাইছে?

Recent Bangladesh full-stack postings cross-check করলে day-to-day core হিসেবে repeatedly আসে JavaScript/TypeScript, React/Next.js, Node.js, REST APIs, SQL/relational database, Git, problem-solving and clear English. More senior roles add PostgreSQL optimization, testing, Docker, CI/CD, AWS/cloud, system design, caching/Redis, queues, security, observability and technical ownership.

Core jobs

JS/TS · React · Node · REST · SQL · Git · problem solving.

Stronger full-stack fit

PostgreSQL · tests · Docker · CI/CD · Next/Nest · cloud.

Senior/AI-era fit

System design · security · observability · architecture · AI-assisted workflow.

AI market signal: Some current Bangladesh roles explicitly mention AI-assisted development tools and LLM integrations. But Stack Overflow 2025 found more developers distrust AI accuracy than trust it, so verification/debugging remains a professional skill.

04 · First professional habit

Requirements & SDLC — coding-এর আগে কী বানাচ্ছ বুঝো

What is it? Requirement = user/business actually কী needs. SDLC = problem থেকে production এবং maintenance পর্যন্ত software-এর life cycle.

Example: “Booking page বানাও” is not enough. Ask: Who can book? Can two users choose the same slot? Payment fail করলে? Cancel করলে? Owner কে notification যাবে?

Must learn MUST PRACTICAL

Functional requirement: system কী করবে.
Non-functional requirement: speed, security, availability, scale, cost.
Acceptance criteria + main edge cases লিখতে পারি.
SDLC flow বুঝি: requirement → design → build → test → release → operate → maintain.

ENOUGH FOR NOW Agile/Scrum theory deep না. Sprint, backlog, ticket, stand-up, review—basic work meaning জানলেই হবে.

Next step when: A feature request শুনে coding-এর আগে important questions এবং acceptance criteria বের করতে পারো.

05 · Foundation

Web & Computer Basics — framework-এর নিচে কী হচ্ছে

Why? Production bug অনেক সময় React/Node problem না; DNS, HTTP, cookies, network, process, environment বা browser problem.

HTTP & browser MUST DEEP BASICS

HTTP

Browser/server কীভাবে request-response exchange করে.

Know: methods, status codes, headers, cookies, caching.

HTTPS / TLS

Network-এর ওপর encrypted connection.

Know: certificate কী role play করে; deep cryptography না.

DNS

Domain name-কে server IP/location-এর সাথে connect করে.

Know: A/CNAME basics and why DNS issue can break a site.

CORS

Browser কোন origin থেকে request allow করবে তার browser-side policy.

Know: CORS auth/security replacement না.

Computer / Linux basics WORKING

Process, CPU, memory, file/permission, environment variable.
Useful commands: ls, cd, cat, grep, tail, ps, kill, curl, ssh.
Browser DevTools: Network, Console, Performance basics.

Next step when: URL type করা থেকে request server-এ গিয়ে response আসা পর্যন্ত basic path explain করতে পারো.

06 · Main language

JavaScript → TypeScript

JavaScript MUST DEEP

React এবং Node দুটোর language foundation. Framework শেখার আগে language behavior clear থাকা দরকার.

Scope, closure, hoisting, this, objects/arrays, mutation vs reference.
Promise, async/await, error handling, event loop, microtask/task.
Modules, destructuring, spread/rest, map/filter/reduce.

TypeScript MUST DEEP

TypeScript তোমার code-এর contracts clear করে. Large app এবং AI-generated code review দুটোর জন্য useful.

Types/interfaces, union, narrowing, type guards, unknown vs any.
Generics and useful utilities: Pick, Omit, Partial, Record.
API/domain/form/component typing with strict: true.

DSA minimum WORKING

Big-O basics, array, hash map/object, Set/Map, stack/queue, binary search, tree/graph BFS/DFS basics. Competitive programming এখন দরকার নেই.

Next step when: Async JS explain করতে পারো and a strict TypeScript feature without any-spam build করতে পারো.

07 · Daily engineering

Git & Debugging

Git MUST PRACTICAL

Git শুধু push করার tool না; team change history, review and safe rollback-এর foundation.

Branch, merge, conflict, rebase basics, revert, stash, cherry-pick basics.
Pull Request, code review, small commits, protected branches concept.
Lockfile, environment config and dependency update discipline.

Debugging MUST DEEP

Reproduce
→ Gather evidence
→ Narrow the problem
→ Form a hypothesis
→ Test the hypothesis
→ Fix root cause
→ Add regression test
→ Verify

Evidence means: stack trace, browser Network tab, server logs, database query/data, request IDs—not repeated guessing.

Next step when: AI wrong suggestion দিলে নিজে evidence দিয়ে root cause বের করতে পারো.

08 · Build UI

Frontend Engineering — React + Next.js

Goal: শুধু সুন্দর UI না. Correct state, accessible forms, predictable data flow, good loading/error states and good performance.

React MUST DEEP

Components, props/state, composition, controlled forms.
useState/useEffect/useRef/useMemo/useCallback/useContext.
Render/re-render, effect cleanup, race conditions.
Server state/caching/mutations — TanStack Query concept.

Next.js STRONG

App Router, layouts, dynamic routes, loading/error/not-found.
Server vs Client Components — কোন code কোথায় run করে.
Data fetching, caching/revalidation, route handlers, cookies/auth flow.

HTML/CSS + Accessibility + Performance MUST PRACTICAL

React-er age basic UI foundation strong korte hobe: semantic HTML, forms, labels, buttons/links, tables, CSS box model, responsive Flex/Grid, media queries, focus states, keyboard navigation, lazy loading, code splitting, image optimization and Core Web Vitals basics.

Real UI states you must build

Loading, empty, error, success, permission denied and validation states.
Search, filter, sort, pagination and detail/edit flows.
Mobile-first responsive layout without horizontal overflow.
Accessible forms: label, error text, focus order and keyboard submit.

Next step when: Production-style frontend বানাতে পারো এবং re-render/network/accessibility problem debug করতে পারো.

09 · Business logic

Backend & API Engineering — Node.js, Express/NestJS, Auth

What is backend? UI-এর পেছনে business rules, authorization, data access, integrations and server-side processing.

Node.js MUST DEEP

Event loop/async I/O, HTTP lifecycle, environment/process, error handling and graceful shutdown basics.

Backend structure MUST DEEP

Route
→ Controller  (HTTP input/output)
→ Service     (business rule)
→ Repository  (data access)
→ Database / External API

This is not a law. Main idea হলো responsibilities mix না করা.

Express + NestJS

Express fundamentals strong রাখবে. NestJS-এ modules, DI/providers, DTO, pipes, guards, interceptors, exception filters working/strong level.

REST API MUST DEEP

Resource design, methods/status codes, validation, errors.
Pagination, filter, sort, search, versioning basics.
OpenAPI/Swagger, webhooks, retries and idempotency concept.

Authentication & Authorization MUST DEEP

Session vs JWT, access/refresh token, cookies, password hashing, OAuth/OIDC basics, RBAC/permissions. Authorization must be server-side.

Next step when: A validated, documented, secured REST API with proper authorization and error handling build করতে পারো.

10 · Data

Database & Data — SQL + PostgreSQL first

Why PostgreSQL/SQL deep? CRM, HRMS, POS, booking, billing-এর মতো business apps-এ relationships, constraints, transactions and reporting খুব important.

PostgreSQL / SQL MUST DEEP

SELECT/INSERT/UPDATE/DELETE, joins, GROUP BY, CTE/subquery, aggregate/window basics.
PK/FK/unique/check/not-null, 1:1/1:N/N:N and normalization.
Transactions, ACID, isolation basics, locks and race conditions.
Indexes/composite indexes and EXPLAIN ANALYZE.
Migrations, backups, restore and safe data changes.

ORM MUST PRACTICAL

Prisma/Mongoose speed দেয়. কিন্তু ORM-এর নিচে data/query কী হচ্ছে বুঝতে হবে.

MongoDB WORKING

Document modeling, indexes, aggregation basics and when NoSQL fits. PostgreSQL-এর আগে Mongo deep study করার দরকার নেই.

Database exercises

Draw ERD for users, teams, customers, deals/tasks and audit logs.
Write migrations with rollback plan and seed data.
Prevent duplicate booking/payment with constraints and transactions.
Compare slow query before/after index using EXPLAIN ANALYZE.
Practice backup restore locally, not only backup creation.

Next step when: Schema design, transaction-based correctness and a slow-query investigation independently করতে পারো.

11 · Maintainable code

Software Design — code যেন change করা safe হয়

What is it? Software Design = class/file বানানোর style না; responsibilities, dependencies and boundaries এমনভাবে arrange করা যাতে code বুঝা, test করা এবং change করা সহজ হয়.

Separation of Concerns

UI, business rule, database logic এক জায়গায় mix না.

High Cohesion

Related responsibilities একই module-এ থাকে.

Low Coupling

এক module change করলে সব module break না.

Simple Interfaces

Modules clearly communicate কী input নেয়, কী output দেয়.

WORKING SOLID/design patterns-এর meaning বুঝবে, কিন্তু acronym/pattern force করবে না. Simple solution first.

Code review must check

Correctness · readability · scope · tests · security · performance · backwards compatibility.

Next step when: Large feature-কে clean modules-এ divide করতে পারো এবং কেন সেই boundary choose করেছ explain করতে পারো.

12 · Prove correctness

Testing & Software Quality

Why? AI বা human—যেই code লিখুক, production-এর আগে evidence দরকার যে important behavior ঠিক আছে.

Testing MUST DEEP

Unit Test

একটা ছোট business function/rule independently test.

Integration Test

API + DB/auth-এর মতো multiple parts together test.

UI Test

User behavior from React Testing Library.

E2E Test

Real critical flow, e.g. login → create → update → logout.

Tools: Vitest/Jest, React Testing Library, Supertest, Playwright. Tool name-এর চেয়ে what to test বেশি important.

Quality is more than tests

Correctness, maintainability, security, performance, accessibility, reliability and operability.

Next step when: Critical feature-এর happy path + edge/error + authorization behavior automated tests-এ cover করতে পারো.

13 · Secure development

Security — feature build করার সময় থেকেই

NIST SSDF secure practices-কে SDLC-এর মধ্যে integrate করতে বলে; OWASP Top 10:2025 web developers-এর major application risks-এর current awareness baseline.

Must know MUST PRACTICAL

Broken access control / IDOR: user যেন অন্য user-এর resource access না করে.
Authentication failures: password/reset/session/token safely handle.
Injection: validate input; parameterized queries/ORM safely use.
Security misconfiguration: CORS, headers, debug mode, public buckets, default secrets.
Supply-chain risk: lockfile, dependency review, minimal packages.
Secrets: source/client/log-এ secret না.
Logging/alerts: security-relevant events visible, sensitive data hidden.
Simple threat modeling: what are we protecting → who can abuse → how → control.

Security exercises

Try to access another tenant/user record by changing an ID in the URL or API request.
Write tests for forbidden access, expired token and missing permission.
Run dependency audit and remove unused packages.
Check that secrets are not in client bundle, Git history, logs or screenshots.
Validate every external input: body, query, params, file upload and webhook payload.

Next step when: Authentication, authorization, validation, secrets and dependency risk-এর basic security review নিজে করতে পারো.

14 · Fast & dependable

Performance & Reliability

Performance MUST PRACTICAL

Guess করে optimize না. First measure.

Frontend: network waterfall, bundle, images, unnecessary renders, Core Web Vitals basics.
Backend: response timing, N+1 queries, slow external API, blocking work.
Database: indexes, query plan, pagination, connection pool concept.

Reliability MUST PRACTICAL

Timeout

External call forever wait করবে না.

Retry

Temporary failure-এ controlled retry; সব operation blindly retry না.

Idempotency

Same request আবার এলেও duplicate payment/action না হয়.

Backup + Restore

Backup আছে বললেই enough না; restore করতে পারা লাগে.

15 · Delivery

Docker, CI/CD & Linux — code থেকে production

Docker MUST PRACTICAL

Image, container, Dockerfile, ports, env, volumes, networks, Docker Compose. Goal: same application reliably different environments-এ run.

CI/CD MUST PRACTICAL

Pull Request
→ Lint + Typecheck
→ Tests
→ Build
→ Staging
→ Smoke/E2E
→ Production
→ Monitor
→ Rollback if needed

Linux WORKING

Server logs/process/permissions/env/network inspect করার মতো knowledge. Linux admin expert না.

Beginner deployment path

Run locally with env vars
→ Add Docker Compose for app + PostgreSQL + Redis
→ CI: lint, typecheck, test, build
→ Deploy one clear path first: VPS, ECS, Render, Railway or similar
→ Add logs, health check and backup
→ Practice rollback after a bad release

Early goal হলো one path reliable করা. একসাথে সব cloud/service শিখতে গেলে depth কমে যাবে.

Next step when: Dockerized app CI দিয়ে build/test করতে এবং failed deployment rollback করতে পারো.

16 · Production ownership

AWS & Observability

AWS minimum MUST PRACTICAL

IAM

Who/what can access AWS resources. Learn least privilege.

EC2 / ECS

Where application compute runs. One path strong করলেই enough initially.

RDS

Managed PostgreSQL database, backups/security basics.

S3

Files/object storage; signed access concept.

CloudFront

CDN/cache static content near users.

CloudWatch

Logs, metrics, alarms.

Observability — system-এর ভেতরে কী হচ্ছে বুঝার ক্ষমতা

Logs

What happened? Error/context/request ID.

Metrics

How much? Request rate, latency, errors, resource use.

Traces

One request কোন service/DB/API path দিয়ে গেল.

Alerts

Important failure হলে engineer-কে জানায়.

AWS Well-Architected architecture review-এ six pillars use করে: operational excellence, security, reliability, performance efficiency, cost optimization, sustainability.

Next step when: AWS-এ app deploy, DB/storage connect, logs/metrics inspect এবং common production issue diagnose করতে পারো.

17 · Whole-system thinking

System Design — পুরো system কীভাবে চলবে?

Beginner definition: System Design মানে “big company diagram” না. It means deciding how the whole application will handle users, data, traffic, slow work, failures and growth.

Example: PlayIndoor-style booking. 100 user থাকলে simple app works. কিন্তু 10,000 user হলে: same slot double-book হবে না কীভাবে? Images কোথায় থাকবে? Search fast কীভাবে? Email send করতে request wait করবে? Server fail করলে?

Concepts you MUST understand

Functional vs Non-functional

Functional: “user can book.” Non-functional: fast, secure, available, scalable.

Load Balancer

Traffic manager. Multiple server instance থাকলে request distribute করে.

CDN

Images/static files user-এর কাছাকাছি server থেকে দেয়, so faster delivery.

Cache / Redis

Frequently needed data-এর fast temporary copy. DB load কমায়. Key question: cache কখন invalid হবে?

Queue / Background Job

Email, PDF, image processing-এর মতো slow work main request থেকে আলাদা queue-তে যায়.

Object Storage

Large files DB row-এ না রেখে S3-type storage-এ রাখা.

Horizontal Scaling

One bigger server না; more app instances add করা.

Consistency / Transaction

Related changes একসাথে correct থাকা. Booking/payment-এ critical.

Idempotency

Same payment/webhook request twice এলে action যেন duplicate না হয়.

Failure Handling

Timeout, retry, backup, graceful degradation. “Dependency fail করলে system কী করবে?”

Basic only for now

BASIC ONLY Replication, sharding, multi-region, eventual consistency—concept কী এবং কেন লাগে জানবে. Hyperscale implementation deep এখন না.

Done when: Booking/CRM system-এর high-level diagram এঁকে DB, cache, queue, storage and failure choices simple ভাষায় justify করতে পারো.

18 · Organize the system

Software Architecture — parts কীভাবে organize হবে?

Beginner definition: Architecture হলো system/code-এর বড় structure. কোন module কী responsibility নেবে, কে কার ওপর depend করবে, data কার ownership—এগুলো decide করা.

System Design vs Architecture — simple difference

System Design

“Whole system under traffic/failure কীভাবে চলবে?”
Example: cache, queue, DB, load balancer.

Software Architecture

“Code/services কীভাবে ভাগ করলে safely change করা যাবে?”
Example: Auth module, Booking module, Payment module.

Architectural styles — how deep?

Modular Monolith — MUST

One deployable application, but clear modules. Small/medium SaaS/business app-এর জন্য strong default.

Layered Architecture — MUST

HTTP/controller → service/business → data layer. Responsibility clear রাখা.

Clean/Hexagonal — WORKING

Core business rules যেন framework/database-এর সাথে unnecessarily tightly coupled না হয়. Idea বুঝবে, dogma না.

Microservices — BASIC/WORKING

Separate services independently deploy. Benefit আছে, but network/ops/data complexity অনেক. Default না.

Event-Driven — BASIC

One event triggers other work, e.g. OrderCreated → email/inventory. Eventual consistency/idempotency basics.

Serverless — BASIC

Cloud functions/events. Useful for some workloads; not a universal architecture.

Documents a professional engineer should create

System diagram: app/API/DB/cache/queue/external systems.
ERD: database entities and relationships.
ADR: important decision — context, decision, trade-off.
Runbook: deploy, migrate, logs, rollback, recovery.

Done when: “Why modular monolith?” বা “Why not microservices?” business/team/complexity trade-off দিয়ে answer করতে পারো.

19 · AI era

AI-Assisted Engineering — AI use করবে, control হারাবে না

DORA 2025 AI-কে an amplifier বলে: strong engineering practice থাকলে benefit বাড়ে, weak process থাকলে problem-ও amplify হতে পারে. Stack Overflow 2025-এ 46% developers AI accuracy distrust করে vs 33% trust.

Professional AI workflow MUST PRACTICAL

Understand task
→ Inspect current code
→ Define acceptance criteria
→ Ask AI for a focused change
→ Review diff
→ Lint + Typecheck + Tests
→ Security / data / performance review
→ Manual verification
→ Commit
→ Monitor

You must own

Requirements and architecture decisions.
Authorization/data access and migration safety.
AI code/test review and rejecting wrong output.
Root-cause debugging when AI fails.
Company/source-code/secrets/privacy policy.

Golden rule: AI generated code তুমি explain/test/debug করতে না পারলে critical production code হিসেবে accept কোরো না.

20 · Build AI features

AI Application Engineering — কতটুকু দরকার?

WORKING AFTER CORE Full-Stack engineer হিসেবে ML researcher হতে হবে না. Real app-এর মধ্যে reliable AI workflow build করা useful.

LLM API

Model call, context/token/cost basics.

Structured Output

Free text নয়; schema-valid JSON/result use করা.

Tool Calling

Model action suggest করে; app validates permission then executes.

RAG

Your trusted documents/data retrieve করে answer ground করা.

Evaluation

AI feature “seems good” না; known test cases দিয়ে quality check.

Human Approval

High-impact action model automatically execute না.

AI safety basics

Prompt injection, sensitive data leakage, excessive tool permission, unvalidated output, unbounded cost/rate. NIST also extends SSDF with AI-specific secure-development guidance.

BASIC ONLY NOW Agents, MCP, multi-agent frameworks. Concept জানো; core software engineering weak রেখে এগুলো chase করো না.

21 · Professional daily work

Communication, Documentation, Reviews & Ownership

Current job postings repeatedly ask for problem-solving, written/spoken English, collaboration, code reviews and technical ownership. এগুলো “soft” বলে ignore করলে international role কঠিন হবে.

Must practice MUST PRACTICAL

Clarifying questions ask করি before building.
Technical decision simple English-এ explain করতে পারি.
Code review করি: correctness, risk, tests, readability, scope.
README/setup/API docs/ADR/runbook লিখতে পারি.
Blocker/incident/status clearly communicate করি.

Estimation

Perfect estimate impossible. Break work into smaller tasks, identify unknowns/risks, state assumptions and communicate confidence.

22 · Proof of skill

One Flagship Project — সবকিছু এক project-এ prove

Best fit: Multi-tenant CRM / Business Operations SaaS. 20 tutorial projects-এর চেয়ে একটি engineered project stronger.

Build up in 3 project levels

Level 1: Fundamentals

Static responsive site + form validation + GitHub README. Proves HTML/CSS/JS basics.

Level 2: Full-stack CRUD

Auth, REST API, PostgreSQL, search/filter/pagination, tests and deploy.

Level 3: Flagship SaaS

Multi-tenant CRM with RBAC, jobs, cache, CI/CD, observability, docs and optional AI.

Do not skip levels: Level 3 শুরু করার আগে Level 2 project independently build, debug and deploy করতে পারা উচিত.

Next.js + TypeScript
        ↓
Node / NestJS API
        ↓
PostgreSQL
        ↓
Redis (cache / queue)
        ↓
Worker (email / report / AI job)
        ↓
S3 / External APIs
        ↓
Docker + CI/CD + AWS
        ↓
Logs / Metrics / Error Tracking

Across everything:
RBAC + Security + Tests + Documentation

Must prove

Portfolio proof

Live demo link, GitHub repo and short product walkthrough video.
README with setup, env vars, architecture, API docs, test commands and deployment notes.
Case study: problem, users, constraints, trade-offs, screenshots and lessons learned.
Known limitations and next improvements. Honest scope looks more professional than fake completeness.

Flagship project rubric

Area	Good evidence	Weak signal
Requirements	Clear users, roles, flows, edge cases and acceptance criteria.	Only feature names, no business rules.
Frontend UX	Responsive, accessible, complete loading/empty/error/form states.	Looks okay only on desktop; broken mobile/forms.
Backend	Validated APIs, clean service boundaries, useful errors and API docs.	Controller has everything; random status codes.
Database	ERD, constraints, migrations, transactions, indexes and restore practice.	Tables exist but data integrity depends on luck.
Security	Server-side RBAC, tenant isolation tests, secret discipline, dependency review.	Auth only hides buttons in UI.
Testing	Unit, integration and at least a few E2E tests for critical paths.	No tests or only happy-path snapshots.
Production	Docker/CI/deploy/logs/backup/rollback documented and tested.	Runs only on your machine.
AI feature	Schema output, validation, evaluation cases and human approval for risky actions.	Raw prompt response directly trusted.
Communication	README, architecture diagram, ADR, runbook and concise case study.	No explanation of decisions or trade-offs.

23 · Scope control

কোনগুলো এখন শুধু basic, deep না

Topic	Level now	Enough for now
Kubernetes	BASIC	Pod/deployment/service কেন লাগে. Deep cluster ops later.
Terraform / IaC	BASIC → LATER	Infrastructure as Code purpose; practical after AWS foundation.
Kafka / RabbitMQ	BASIC	Message broker/queue purpose, retry/delivery basics.
GraphQL	BASIC	Schema/query/mutation and when it may fit better than REST.
Microservices	CONCEPT	Trade-offs; do not make it default.
Python/FastAPI	LATER	Useful second language for AI/data; TypeScript remains your core.
Go / Rust / Java / .NET	NOT NOW	Learn if repeated target jobs/project requires it.
Advanced ML / PyTorch	NOT NOW	Not required for Full-Stack + AI integration path.
MCP / multi-agent frameworks	BASIC	Know the idea; deep only when work needs it.

24 · Practical calendar

32-week focused plan

Week

Focus

Output

Depth

1

Requirements + Web basics

Feature spec + request-flow notes.

Foundation

2

HTML/CSS/A11y

Responsive page with accessible form and mobile QA.

Foundation

3–4

JS + TypeScript

Strict TS mini-module + async debugging.

Deep

5

Git + Debugging

PR workflow + root-cause exercise.

Practical

6–8

React + Next.js

Production-style frontend module with all UI states.

Deep

9–12

Node/Nest + APIs/Auth

Validated, documented secured API.

Deep

13–15

PostgreSQL

ERD + schema + transaction + index/query exercise.

Deep

16–19

Design + Tests + Security

Refactor module + test suite + security abuse tests.

Deep

20–23

Docker + CI/CD + Deploy

One reliable deployment path, logs, backup and rollback.

Practical

24–26

Observability + Reliability

Logs/metrics + incident/debug exercise.

Practical

27–29

System Design + Architecture

3 system diagrams + 2 ADRs.

Deep basics

30–31

AI Engineering

AI workflow + reviewed AI feature.

Working

32

Polish & Interview

Flagship docs/demo + interview explanations.

Proof

Do not stop applying: Learning and job applications run in parallel. Job descriptions can change weekly priority, but not your core foundation.

25 · Self-check

Professional Full-Stack Engineer readiness checklist

Good target: Most answers “yes” + flagship project has evidence = strong professional direction. You do not need to know every tool in the ecosystem.

Interview and application readiness

Resume should show outcomes, stack, ownership and links. শুধু technology list না.
GitHub pinned project should have clear README, screenshots, setup, env example and test/deploy commands.
Practice explaining one feature end-to-end: requirement, API, DB, UI, tests, security and deployment.
Practice take-home discipline: clarify scope, make small commits, write notes, handle edge cases and submit clean instructions.
Prepare stories for debugging, trade-off, conflict, missed estimate and production issue.

26 · Avoid these traps

Common beginner mistakes

Tutorial copy-paste

Course finish করলে skill prove হয় না. Same idea দিয়ে new feature independently build করতে হবে.

CSS and accessibility ignore করা

Full-stack engineer হলেও broken responsive UI, missing labels and bad forms professional signal কমায়.

Database weak রাখা

Real business apps data correctness-এর ওপর দাঁড়ায়. Constraints, transactions and migrations skip করা যাবে না.

Client-side auth trust করা

Button hide করা authorization না. Permission check server-side হতে হবে.

No tests, no verification

“Works on my machine” enough না. Important paths test, lint, typecheck and manually verify করতে হবে.

Deploy না করা

Localhost project job proof না. Deploy, logs, env, backup and rollback practice দরকার.

Too many tools chase করা

Kubernetes, microservices, new language, AI agents সব একসাথে শিখলে core weak থাকে. Need-based learning করো.

AI blindly trust করা

AI output explain/test/debug করতে না পারলে production-level code হিসেবে accept করবে না.

27 · Authentic basis

Research sources used

Reviewed 28 Aug 2026. Professional standards define the foundation; current job postings are used only to cross-check market demand.

Trusted learning resources

Topic	Best starting resource	Use it for
HTML/CSS/JS	MDN Web Docs	Web platform fundamentals, forms, accessibility, HTTP, browser APIs.
Accessibility	W3C WAI Fundamentals	Semantic HTML, keyboard/focus, labels and inclusive UI basics.
React	React Learn	Components, state, effects, rendering and modern React mental model.
Next.js	Next.js Docs	App Router, server/client components, routing, caching and deployment basics.
TypeScript	TypeScript Docs	Types, narrowing, generics, strict mode and app-level contracts.
Node.js	Node.js Learn	Runtime, modules, async behavior, HTTP/server basics.
NestJS	NestJS Docs	Modules, providers, DTOs, guards, pipes, testing and backend structure.
PostgreSQL	PostgreSQL Docs	SQL, constraints, indexes, transactions, query plans, backup/restore.
Testing	Vitest Guide + Playwright Docs	Unit/integration tests and realistic browser E2E tests.
Security	OWASP Top 10 + OWASP Cheat Sheets	Common web risks and practical secure implementation checks.
Docker	Docker Get Started	Images, containers, Dockerfile, Compose and local production-like setup.
AWS	AWS Well-Architected	Production thinking: security, reliability, operations, cost and performance.
AI Apps	OpenAI Platform Docs + OWASP GenAI Security	LLM APIs, structured output, tool use, evaluation and AI security risks.

IEEE Computer Society — SWEBOK Guide V4.0 / V4.0aDefines the recognized software-engineering knowledge areas and the discipline’s professional scope.
https://www.computer.org/education/bodies-of-knowledge/software-engineering/topics

NIST SP 800-218 — Secure Software Development Framework (SSDF) 1.1Secure software-development practices integrated across the SDLC.
https://csrc.nist.gov/pubs/sp/800/218/final

OWASP Top 10:2025Current web-application security awareness baseline.
https://owasp.org/Top10/

AWS Well-Architected FrameworkOperational excellence, security, reliability, performance efficiency, cost optimization and sustainability.
AWS Well-Architected

DORA — State of AI-assisted Software Development 2025AI is primarily an amplifier of existing organizational/engineering strengths and weaknesses.
https://dora.dev/research/2025/dora-report/

Stack Overflow Developer Survey 2025 — AI46% distrust AI accuracy vs 33% trust; “almost right” solutions and AI-code debugging remain common frustrations.
https://survey.stackoverflow.co/2025/ai

Current Bangladesh job-market cross-checks

Bdjobs / R.E.S. LLC — Full-Stack TypeScript DeveloperStrong JS/TS, Node, React, REST, SQL, Git, English/problem-solving; PostgreSQL, testing, Docker, CI/CD, NestJS/AWS as stronger-fit skills; LLM/AI-assisted development as bonus.
Job snapshot

Next Solution Lab — Senior Full Stack EngineerReact/Next/TypeScript, Node/Nest, PostgreSQL, Redis, AWS, CI/CD, Docker/Kubernetes, system design, queues, security and observability.
Job snapshot

MVI Solutions — Senior Full Stack DeveloperREST APIs, relational DB optimization, performance/scalability/caching/async processing, cloud, Git, CI/CD, automated testing, Docker and technical ownership.
Job snapshot

Inovetix — Senior Full Stack EngineerTypeScript, React/Next, Node/Nest, REST, PostgreSQL/MySQL, Git, Docker/CI-CD familiarity and AI-assisted debugging/code-generation experience.
Job snapshot

Final rule: No single job post decides your roadmap. Repeated skills across professional standards + multiple current roles decide the core; one-off technologies stay optional.