Was choosing CSE in 2026 a mistake?
Use this first to separate AI anxiety from the longer decision to learn computing well.
Published title: ২০২৬ এ এসে CSE তে ভর্তি হওয়া কি ভুল সিদ্ধান্ত।Computer Science। Jhankar Mahbub
Quality over quantity
Not a resource directory. Curated starting points.
Resources work best in roadmap context. This index lets students and maintainers inspect selection, scope and review dates in one place.
Bangladesh tech leaders & practitioners
Not a random feed—a deliberate watch order.
These nine videos and podcast episodes first settle the anxiety, then strengthen thinking, and finally widen the view of industry. They are perspectives—not final verdicts, endorsements or job guarantees.
1 · Ground yourselfStart with the decision you already made and the habits that can still change. Do not begin with salary or panic.2
Student-life red flags that quietly cost semesters
Turn motivation into an audit of time, habits and avoidable drift—not guilt or comparison.
Published title: যে ভুলগুলো করলে ইউনিভার্সিটি জীবন ধ্বংস হবেই !। Red Flags in Student Life । Jhankar Mahbub
2 · Strengthen how you thinkThen examine why programming, problem solving and engineering judgment still matter when AI can produce code.3
Why learn programming when AI can code?
Connect programming practice to understanding, verification and the ability to direct tools responsibly.
Published title: AI আসছে তবুও কেন প্রোগ্রামিং শিখবে?
Why many learners get stuck in problem solving
Look for a better practice loop; treat the percentage in the title as presentation, not a statistic supplied by CSE-Pothik.
Published title: যে ভুলের কারণে ৯৫% মানুষ প্রবলেম সলভিং এ ব্যর্থ হয় | Nafis Shahriar | SWE Edition - Phicast
Is knowing how to code enough to get hired?
Compare coding output with problem framing, communication and the judgment expected in engineering work.
Published title: কোডিং জানলেও চাকরি জুটবে না? | Hasan Shahriar Masud
Use the pattern
Learn → Practice → Go deeper (optional)
Finish the recommended scope of one primary resource, build something, then decide whether deeper material is needed.
25 curated resources
Official documentation
Maintained primary references for tools, platforms and standards.
Wireshark User’s Guide
Official, free guidance for inspecting packet evidence and understanding the tool’s limits.
Do this much: Read capture prerequisites, filtering and following streams. Use only your own isolated lab or an explicitly authorized sample; captures may contain secrets.
Last reviewed:
Getting Started in KiCad
Connects an editable circuit design with checks and fabrication outputs without requiring a board purchase.
Do this much: Use the tutorial matching your installed version: schematic, electrical rules and board layout. Software/docs are free; components and fabrication are not included.
Last reviewed:
Writing tests with Playwright
Teaches executable checks and observable outcomes without treating automation as the whole of QA.
Do this much: Write one isolated user-journey test with meaningful locators and assertions; then test a failure state. Use the free local runner, not an optional hosted service.
Last reviewed:
Flutter documentation
A maintained starting point for the Flutter branch, with platform-specific setup made explicit.
Do this much: Pick your available target device, check setup and learn Dart/widgets/state for one small app. Paid hosted tools are unnecessary for this exercise.
Last reviewed:
Get started with React Native
Explains the supported starting workflow without pretending cross-platform removes native constraints.
Do this much: Use if you already know JavaScript/React; read the framework and environment choices, then implement one screen. Check optional service costs separately.
Last reviewed:
User research: Service Manual
Connects research questions, ethical participation and evidence instead of invented personas.
Do this much: Read consent and usability-testing guidance; plan one small observation session. This is a learning exercise, not a representative research study.
Last reviewed:
GOV.UK Design System
Shows how research-informed patterns become documented, reusable interfaces.
Do this much: Study a form component and its error pattern. Learn the reasoning; do not copy government branding into your portfolio.
Last reviewed:
MDN Learn Web Development
Standards-first web learning without framework noise.
Do this much: Core HTML, CSS, JavaScript and accessibility modules.
Last reviewed:
Web Accessibility Tutorials
Authoritative guidance that turns accessibility into engineering practice.
Do this much: Start with page structure, menus, images and forms.
Last reviewed:
Node.js Learn
An official path from JavaScript knowledge to server-side reasoning.
Do this much: HTTP, asynchronous work, modules and testing basics.
Last reviewed:
PostgreSQL Tutorial
Learn relational ideas from the database itself, not an ORM abstraction.
Do this much: Tables, queries, joins, aggregates and transactions.
Last reviewed:
Docker Get Started
Enough container literacy for a first deployment—no Kubernetes detour.
Do this much: Containerize one small app and understand images, containers, ports and volumes.
Last reviewed:
Arduino Documentation
Immediate hardware feedback without hiding basic embedded constraints.
Do this much: Digital/analog I/O, sensors and one small timing-based project.
Last reviewed:
Power BI learning path
Official, task-oriented guidance that connects a report to its data model.
Do this much: Use the beginner modules for data preparation, modeling and accessible reports; one BI tool is enough to start.
Last reviewed:
What is Retrieval-Augmented Generation (RAG)?
A maintained conceptual overview that separates RAG architecture from a fashionable vector-database demo.
Do this much: Learn the retrieval → augmentation → generation flow and why data ingestion, retrieval quality, grounding and serving must be reasoned about separately.
Last reviewed:
n8n learning paths
Official tool guidance for implementing visible workflows without confusing a canvas with engineering judgment.
Do this much: Learn workflow structure, data flow, branches, loops, error handling and executions before adding an AI node. Rebuild one small, harmless workflow and inspect its failure history.
Last reviewed:
OpenClaw agent runtime and security model
A concrete runtime example that makes persistent agent authority—and the resulting trust boundary—visible.
Do this much: After building one bounded agent, inspect its loop, workspace, tool wiring and sessions; then read the linked security, sandbox and tool-policy guidance before granting host or account access.
Last reviewed:
Cyber Defense Analyst work role
An official task-level description of defensive analysis work, including alert correlation, documentation and escalation.
Do this much: Read the role purpose, core tasks and knowledge list. Use it to distinguish evidence-based monitoring and escalation from the movie version of a SOC.
Last reviewed:
NIST SP 800-61 Rev. 3 — Incident Response
Current, vendor-neutral incident-response guidance that keeps detection and recovery connected to organizational risk.
Do this much: Start with the executive summary and incident-response life-cycle considerations. Map preparation, detection, response and recovery to one fictional incident; do not treat the publication as a tool tutorial.
Last reviewed:
Get Started with MITRE ATT&CK
A maintained common language for reasoning about observed adversary behavior, detections and defensive gaps.
Do this much: Learn tactics, techniques, procedures and the detection/analytics use case. Map only a few relevant behaviors to available log evidence; never chase 100% matrix coverage.
Last reviewed:
OWASP Web Security Testing Guide v4.2
A vendor-neutral testing methodology that frames penetration testing as scoped investigation and reporting—not indiscriminate scanning.
Do this much: Read the introduction and testing framework, then choose only the versioned scenarios relevant to one explicitly authorized training application. Record scope, evidence and limits.
Last reviewed:
OWASP DevSecOps Guideline
Connects security activities to delivery stages while making clear that tools support—not replace—engineering review.
Do this much: Read the pipeline overview, threat modelling, secrets and scanning sections. Add one reviewed check to a practice pipeline and document false-positive handling and a safe exception path.
Last reviewed:
OWASP Top 10
A shared starting vocabulary for application-security risk, not a complete test plan.
Do this much: Use the categories to recognize common application risks, then verify behavior in an authorized lab rather than scanning public targets.
Last reviewed:
OWASP Cheat Sheet Series
Actionable prevention guidance organized by the problem developers are implementing.
Do this much: Select the sheet matching one feature, such as input validation or authentication; test the recommendation in your own small app.
Last reviewed:
Threat Modeling Cheat Sheet
Moves security earlier than scanner output and keeps it tied to system context.
Do this much: Map one small feature's data flow, trust boundaries, threats and mitigations with its owners.
Last reviewed:
12 curated resources
Courses & university material
Structured learning material for foundations and guided sequences.
Android Basics with Compose
An official learning sequence that connects language basics with real app behavior and tests.
Do this much: Start with Kotlin and the first interactive app; then state and navigation. Choose this branch for Android, not alongside every other mobile stack.
Last reviewed:
Develop in Swift
Official Swift and Apple-platform guidance instead of mixing outdated tutorials.
Do this much: Check tooling requirements first, then build one view and state transition. Tutorials are free; hardware and distribution may have costs.
Last reviewed:
CS50x: Introduction to Computer Science
A rigorous, friendly foundation in problem solving, C, memory and data structures.
Do this much: Weeks 0–5; finish the problem sets you can explain.
Last reviewed:
Data Engineering Zoomcamp
A realistic, project-oriented introduction to pipelines and warehouses.
Do this much: Modules 1–3 first; orchestration and streaming can wait.
Last reviewed:
Machine Learning Crash Course
Concise theory linked to practical model behavior and evaluation.
Do this much: Regression, classification, data and overfitting modules.
Last reviewed:
Linux Journey
A practical foundation before security or cloud tooling.
Do this much: Grasshopper and Journeyman sections; type every command yourself.
Last reviewed:
The Missing Semester of Your CS Education
The everyday engineering tools many curricula leave implicit.
Do this much: Shell, editors, data wrangling, Git and debugging lectures.
Last reviewed:
Nand2Tetris
Builds a concrete mental model of what lies beneath software.
Do this much: Projects 1–5 to connect gates, CPU and machine code.
Last reviewed:
Algorithms, Part I
A rigorous bridge from solving problems to reasoning about algorithms.
Do this much: Audit core data structures, sorting and graph foundations.
Last reviewed:
LLM Course
A free technical foundation that makes model behavior less mysterious.
Do this much: Study model and transformer foundations before using the later application chapters selectively.
Last reviewed:
Full Stack Deep Learning
Connects models to the engineering work required around production systems.
Do this much: Use selected lectures on development infrastructure, testing, deployment and continual learning; verify current tool details separately.
Last reviewed:
Made With ML
Treats evaluation and reliable delivery as part of ML/AI engineering.
Do this much: Focus on data, evaluation, testing and production lessons; reproduce a small example rather than copying the full stack.
Last reviewed:
5 curated resources
Labs & practice
Places to turn reading into small, safe evidence.
SQLBolt
Fast feedback for turning SQL syntax into usable skill.
Do this much: Complete lessons 1–12 and the exercises.
Last reviewed:
Intro to Machine Learning
A low-friction way to train, validate and inspect a small model.
Do this much: Finish the short course and one notebook; skip competitions for now.
Last reviewed:
Web Security Academy
Safe, legal, deliberately vulnerable labs with excellent explanations.
Do this much: Read the getting-started material and complete beginner SQLi/XSS labs only in the academy.
Last reviewed:
Pandas
Quick practice for selecting, cleaning, grouping and checking tabular data.
Do this much: Complete the short exercises, then repeat one analysis on a small CSV you understand.
Last reviewed:
CP-Algorithms
Clear algorithm explanations for the competence and competitive tracks.
Do this much: Use as a reference after attempting problems, not as a list to memorize.
Last reviewed:
14 curated resources
Books & focused articles
Carefully scoped reading for context, deeper reasoning or a focused topic.
An Introduction to Computer Networks
A freely readable university-authored book that explains mechanisms rather than vendor command memorization.
Do this much: Read the overview, Ethernet, IP and transport introductions. Sketch a packet’s route before deeper routing chapters.
Last reviewed:
Foundation Level testing syllabus
Provides shared testing vocabulary; pair it with a real app and evidence rather than exam memorization.
Do this much: Read the freely downloadable syllabus sections on test design, risk and defect management. Training and the optional certification exam can cost money; neither is required here.
Last reviewed:
Pro Git
The official, durable explanation of the tool Future You will wish you learned early.
Do this much: Chapters 1–3 now; branching workflows later.
Last reviewed:
Teach Yourself Computer Science
A curated map for students who want depth beyond tool tutorials.
Do this much: Use the subject map; choose one deep subject, not all at once.
Last reviewed:
The Accidental CTO
A free, story-driven tour of how technical and leadership decisions evolve as a product grows, useful for connecting isolated tools to operational consequences.
Do this much: Read selectively after basic backend and systems foundations. Treat the incident-driven chapters as prompts to compare scaling and reliability trade-offs—not as a universal architecture recipe.
Last reviewed:
Clean Code and Refactoring
Connects readability, simplicity, duplication and tests to the real maintenance cost of code—including code drafted with AI.
Do this much: Read the clean-code characteristics, then improve names, duplication or one oversized function in working code. Protect behavior with tests before a larger refactor.
Last reviewed:
Code Smells
A visual vocabulary for spotting maintainability pressure that compiling and AI-generated confidence do not reveal.
Do this much: Begin with Bloaters, Change Preventers and Couplers. Find one example in your own project, describe its change cost and refactor only with behavior protected.
Last reviewed:
Design Patterns
Explains patterns as adaptable design blueprints and shared team language, while also making room for their limits and misuse.
Do this much: Read what a pattern is, its benefits and criticism. Study Strategy or Adapter only after you can point to that problem in working code; compare it with the simpler option.
Last reviewed:
JavaScript Testing Best Practices
Practical guidance on behavior-focused tests and maintainability.
Do this much: Read testing fundamentals and apply three tests to a small app.
Last reviewed:
OpenIntro Statistics
A rigorous, openly available statistics foundation for honest analysis.
Do this much: Use the free text for distributions, study design, inference and the difference between association and causation.
Last reviewed:
AI Engineering Roadmap
A practical engineer-to-engineer view that ends stages with something built instead of a certificate.
Do this much: Use the staged path to understand prompts, retrieval, evaluation and production concerns; treat dated provider details as changeable.
Last reviewed:
A practical guide to building agents
A primary guide that puts workflow fit, evals, bounded tools and guardrails before multi-agent complexity.
Do this much: Read what qualifies as an agent, when a deterministic workflow is enough, the single-agent loop, tool design, guardrails and human intervention. Treat SDK examples as one implementation, not the concept itself.
Last reviewed:
শাফায়েতের অ্যালগরিদম লেখা
A broad Bangla index of programming, data-structure and algorithm explanations grounded in problem solving.
Do this much: Start with complexity, basic data structures and graph foundations; attempt a problem before reading an advanced explanation.
Last reviewed:
বাংলা programming ও competitive-programming লেখা
Adds a Bangladesh-based practitioner perspective alongside structured algorithm references.
Do this much: Use the Bangla category selectively for competitive-programming motivation, practice reflection and related engineering notes.
Last reviewed: