CS CONNECT
Home About Podcasts Resources Join

Free Resources

Entry test guides, workshop recordings, and stuff that actually helps.

001 / Sessions & Workshops

Learn From the Community

01 VIDEO

The Brutal and Controversial University Guide

Ft. Azan Waseem. An honest look at Pakistani CS universities: rankings, pros, cons, and what students actually go through.

Watch on YouTube →
The Brutal and Controversial University Guide
02 WEBINAR

Prompt Engineering for Medical Students

RMU Webinar. How AI and prompt engineering are changing the medical field, with practical tips for students in any discipline.

Watch on YouTube →
Prompt Engineering for Medical Students | RMU Webinar
002 / Entry Test Prep

University Entry Tests

PREP

Past papers, test patterns, subject guides, and prep strategies for FAST, NUST, COMSATS, GIKI, LUMS, and more. Put together by students who cracked it.

Join Entry Test Group →
003 / Roadmaps

Learning Roadmaps

01

Entry Test Preparation Tips

Universal tips that work for every entry test. ECAT, NET, NU Test, MDCAT, and beyond. Build the right habits before you start grinding.

Phase 1

Know Your Exam

Get to know what the syllabus and pattern of your exam is. How many questions are there? Which subjects? What's the marks distribution? Is there negative marking? Figure all of this out before you start anything.

Phase 2

Create Your Schedule

Create your timetable of what hours to study each day. Make a schedule of 6 days and keep one day, say an "Extra Day" (most likely Sunday) to catch up on what you missed, because maybe you get sick or make a plan, etc.

Phase 3

Start with the Highest Weightage

Start with the subject that has the most weightage. Practice and learn it every day and then slowly add other subjects in your routine.

Phase 4

Use Effective Study Methods

Use study methods like Active Recall, Spaced Repetition, Feynman Technique, and the Blurting Method to keep track of what you learned previously and move it into long-term memory.

Phase 5

Track Your Mistakes

Keep track on a diary or spreadsheet of what topics you have missed or questions you got wrong to practice them on your Extra Day.

Phase 6

Revise & Take Mock Tests

In the last 2 to 3 weeks, practice topics that you find hard and do an overall revision by taking mock tests to find your weak points. For example, if you forget formulas, could not keep up with time, you just did not know that concept (like a new topic), or you made some silly mistakes.

Phase 7

Simulate Real Exam Conditions

If your exam is 3 hours long, sit for 3 hours straight. No phone, no snacks, no music and do it at the exact time of day as the actual exam. For example, if the exam is at 9:00 AM, practice your test at 9:00 AM.

Phase 8

Final Days

In the last 2 to 3 days, do not panic and trust yourself. Do not learn new topics. Only study the summary of every formula or shortcut from a cheat sheet.

Pro Tips

Important Reminders

Do not pull all-nighters by blasting yourself with a high amount of caffeine. Try to fix your sleep schedule instead. Do not fall for the quantity trap. Quality matters more! Like collecting 10 different PDFs and 5 different YouTube playlists for one subject. Information overload is the fastest way to feel overwhelmed.

02

ECAT (UET) Preparation Roadmap

Subject by subject strategy for cracking the UET Engineering College Admission Test. Prioritized by weightage and difficulty.

Phase 1

Priority Order

We should prepare subjects in this order: Mathematics, Physics, Computer Science, and English. Mathematics and Physics need the most time, while English needs the least.

Phase 2

Mathematics (Top Priority)

Mathematics is one of the most important subjects in ECAT. Your 1st year and 2nd year concepts must be clear. Focus especially on Trigonometry, Algebra, Probability, Conic Sections, Derivatives, and Integration. Understand concepts and formulas properly, solve textbook examples, practice MCQs from entry test books, and solve past ECAT papers. The more MCQs you practice, the faster you will solve questions in the test.

Phase 3

Physics (Second Priority)

Physics is also very important because many questions are concept based. Make sure your 1st year and 2nd year concepts are clear. Understand formulas and when to use them, practice MCQs regularly, solve past ECAT papers, and watch revision marathons before the exam.

Phase 4

Computer Science (Third Priority)

Computer Science mainly depends on ICS syllabus, especially C++. Focus on data types and variables, operators, if else and switch statements, loops, arrays, and functions. Many MCQs are based on C++ code logic or output, so understanding how code works is important. Practice MCQs and past ECAT questions.

Phase 5

English (Lowest Priority)

English does not require much time. Most questions are comprehension based, so you just need basic grammar knowledge. Learn basic grammar rules, practice reading comprehension passages, and solve past MCQs. Even 10 to 15 minutes of daily practice is enough.

Phase 6

Important Tips

ECAT currently has no negative marking, so try to attempt every question. Try to appear in both ECAT attempts (ECAT-1 and ECAT-2) to get the highest possible score.

03

FAST NU Test Roadmap

Complete strategy to ace the FAST National University entrance exam. Covers paper pattern, subject priorities, and mock test practice.

Phase 1

Understand the NU Test and Its Requirements

Gather all your information on the NU Test like what are its dates, what is the paper pattern, and its syllabus. NU Test is conducted once each year, usually around July. It has a total of 120 MCQs. The paper pattern is: Advanced Mathematics has 50 MCQs (1 mark each), Basic Mathematics has 20 (1 mark each), IQ has 20 (1 mark each), and English has 30 (0.33 marks each). Total marks are 100. The syllabus of Advanced Mathematics is from Punjab Textbooks of intermediate. FAST has negative marking too, with ΒΌ deduction for each wrong answer.

Phase 2

Create a Realistic Study Timetable

Create a study schedule that clearly defines at what time and how many hours you will study each day. Plan your schedule for six days of the week and keep one day (most likely Sunday) as an "Extra Day". Use this day to catch up on missed topics or whatever you could not complete because of illness or unexpected plans. This keeps your preparation consistent.

Phase 3

Smart Prioritization

As seen in the paper pattern, your advanced maths carries more weightage. If you have scored 80+ in NET, consider it done for the NU Test too, otherwise watch short tricks of Sir Hashim Zia and practice from KETS (KIPS Books). Keep practicing and topics like Sequences, Trigonometry, Differentiation, Integration, and Conics matter more. Basic maths is usually matric level maths which can be practiced from the Quantitative and Analytical Reasoning section of KETS. IQ can also be covered from the KETS IQ portion and past paper practice questions.

Phase 4

Use Effective Study Techniques

Use study methods such as Active Recall, Spaced Repetition, the Feynman Technique, and the Blurting Method to reinforce your understanding. These techniques help ensure that what you study stays in your long-term memory.

Phase 5

Track Your Mistakes

Keep track of the topics you missed or the questions you answered incorrectly in a diary or spreadsheet. This record will help you identify weak areas and revisit them on your Extra Day, ensuring that mistakes are corrected and important concepts are strengthened.

Phase 6

Plan the Last 2 to 3 Weeks Strategically

In the final 2 to 3 weeks before the FAST admission test, focus on improving weaker topics and your problem solving speed. Take regular mock tests to see where you stand and find issues like time pressure, forgotten formulas, unfamiliar concepts, or silly mistakes. Use these insights to adjust your preparation accordingly.

Phase 7

Practice with Mock Tests

When practicing mock tests, try to make it like the real exam. The NU Test lasts two hours, so sit and attempt the test for the full two hours without distractions. No phone, no snacks, no music. Also try to take the mock test at the same time as your actual exam. Like if your exam is at 9:00 AM, practice your test at 9:00 AM as well. If a laptop is available, take your mock test at fastmock.vercel.app to get a better grip on the computerized test system. If you do not know the answer, leave the question blank because there is negative marking in the NU Test.

Phase 8

Final 2 to 3 Days: Revise Only

In the final 2 to 3 days before the exam, stay calm and trust your preparation. Do not attempt to learn new topics. Focus only on revising summaries, formulas, and shortcuts from your cheat sheet and practice those to boost confidence.

04

NUST NET Roadmap

Full strategy to crack the NUST Entry Test. Covers paper pattern, subject priorities, and how to make the most of all 4 NET attempts.

Phase 1

Understand the NET and Its Requirements

Gather all your information on NET like what are its dates, what is the paper pattern, and its syllabus. NET is conducted 4 times every year, the first 3 are before inter board exams and the last one is after board exams. NET (Engineering) has a total of 200 MCQs each carrying 1 mark. The pattern is: Mathematics has 100 MCQs, Physics has 60 MCQs, and English has 40 MCQs. Total marks of NET (Engineering) are 200. The syllabus is from Punjab and Federal Textbooks of intermediate.

Phase 2

Create a Realistic Study Timetable

Create a study schedule that clearly defines at what time and how many hours you will study each day. Plan your schedule for six days of the week and keep one day (most likely Sunday) as an "Extra Day". Use this day to catch up on missed topics or incomplete tasks due to illness or unexpected plans, ensuring your preparation stays consistent.

Phase 3

Smart Prioritization

NET Mathematics is challenging and often consumes the most time during the exam. Practice at least 50 MCQs daily to aim for a score of 80+ and learn shortcuts from Sir Hashim Zia's lectures. Save time in English by learning techniques for synonyms. NET Physics is largely theory-based, so thorough book reading is essential to score 40+, however practicing numerical MCQs can further improve your performance. As there are 100 MCQs of maths and the test will be computerized, no such topic is "more important". MCQs from each chapter will be asked.

Phase 4

Use Effective Study Techniques

Use study methods such as Active Recall, Spaced Repetition, the Feynman Technique, and the Blurting Method to reinforce your understanding. These techniques help ensure that what you study stays in your long-term memory.

Phase 5

Track Your Mistakes

Keep track of the topics you missed or the questions you answered incorrectly in a diary or spreadsheet. This record will help you identify your weak areas and revisit them on your Extra Day, ensuring that mistakes are corrected and important concepts are strengthened.

Phase 6

Plan the Last 2 to 3 Weeks Strategically

In the last 2 to 3 weeks before the NUST Entrance Test (NET), focus on practicing the topics you find most difficult. Regularly attempt full-length mock tests to evaluate your preparation and identify weak areas such as forgotten formulas, poor time management, unfamiliar concepts, or careless mistakes.

Phase 7

Practice with Mock Tests

When practicing mock tests, try to simulate the real exam environment. NET lasts three hours, so sit and attempt the test for the full three hours without distractions. No phone, no snacks, no music. Also try to take the mock test at the same time as your actual exam. For example, if your exam is at 9:00 AM, begin your practice test at 9:00 AM as well.

Phase 8

Final 2 to 3 Days: Revise Only

In the final 2 to 3 days before the exam, stay calm and trust your preparation. Do not attempt to learn new topics. Focus only on revising summaries, formulas, and shortcuts from your cheat sheet and practice those to boost confidence.

Pro Tips

Things Most People Get Wrong

Do not pull all-nighters and avoid a high amount of caffeine to ensure your mind stays fresh and alert for exam day. Also, do not fall for the quantity trap. Quality matters more. Collecting 10 different PDFs and 5 different YouTube playlists for one subject is not preparation. Information overload is the fastest way to feel overwhelmed.

05

Web Development Roadmap

From your first HTML tag to deploying full-stack apps. The most in-demand skill for Pakistani tech jobs and freelancing.

Phase 1

HTML, CSS & How the Web Works

Learn HTML (semantic tags, forms, accessibility) and CSS (box model, Flexbox, Grid, responsive design). Build 3-5 static pages: a portfolio site, a restaurant landing page, a clone of a site you like. Use MDN Web Docs as your reference, and follow The Odin Project or freeCodeCamp for structure. Don't skip responsive design; most Pakistani users browse on phones.

Phase 2

JavaScript & DOM

Get solid on JavaScript: variables, functions, arrays, objects, DOM manipulation, events, and async/await. Build interactive projects: a to-do list, a quiz app, a weather app using a public API. javascript.info is the best free resource. Learn Git & GitHub here too; you'll need it for everything that follows.

Phase 3

Frontend Framework

Pick React (most jobs ask for it) or Next.js (React with server-side rendering, routing built in). Learn components, state, hooks, and routing. Build something real: a blog with markdown, a dashboard, or an e-commerce frontend. The official React docs at react.dev are excellent now.

Phase 4

Backend & Databases

Learn Node.js + Express or go the Python route with Django/FastAPI. Understand REST APIs, authentication (JWT, sessions), and CRUD. Pick PostgreSQL for relational data or MongoDB for document-based. Build a full-stack project with user auth. Neon or Supabase give you free hosted Postgres to start with.

Phase 5

Deploy, Ship & Get Hired

Deploy on Vercel or Railway. Learn basic CI/CD. Build a portfolio with 3+ projects and push everything to GitHub. Start applying for internships (check LinkedIn, Rozee.pk, and the CS Connect job channel). Pakistani companies hiring web devs include Systems Limited, Arbisoft, 10Pearls, and dozens of startups on LinkedIn.

06

Machine Learning, Deep Learning & CV Roadmap

The path from Python basics to training your own models. Covers classical ML, deep learning, computer vision, and reinforcement learning.

Phase 1

Python & Math Foundations

Get comfortable with Python (you probably already know it from uni). Learn NumPy, Pandas, and Matplotlib for data manipulation and visualization. Brush up on linear algebra (vectors, matrices, eigenvalues) and probability/statistics (distributions, Bayes' theorem). 3Blue1Brown on YouTube makes the math intuitive. Khan Academy works for stats.

Phase 2

Classical Machine Learning

Learn regression, classification, clustering, and decision trees using scikit-learn. Understand train/test splits, cross-validation, bias-variance tradeoff, and feature engineering. Do projects on Kaggle to get real with messy datasets. Andrew Ng's Machine Learning Specialization on Coursera is the gold standard here and often available for free with financial aid.

Phase 3

Deep Learning & Neural Networks

Learn neural network fundamentals: perceptrons, backprop, activation functions, optimizers. Use PyTorch (industry standard now) or TensorFlow. Build CNNs for image classification and RNNs/Transformers for text. fast.ai is the best practical course, gets you training models in week 1. Andrej Karpathy's YouTube series on neural nets is phenomenal for understanding what's actually happening under the hood.

Phase 4

Computer Vision & Reinforcement Learning

For CV: learn object detection (YOLO, Faster R-CNN), image segmentation, and GANs. OpenCV for classical CV tasks. For RL: understand MDPs, Q-learning, policy gradients, and PPO. Spinning Up by OpenAI is a great RL resource. Build projects: a face detection system, a self-playing game agent, or an image style transfer tool. Use Google Colab for free GPU access.

Phase 5

Research & Career

Read papers on arXiv and Papers With Code. Try reproducing results from interesting papers. Look into ML roles at Pakistani companies like Ailaaj, Afiniti, or research labs at NUST/LUMS/ITU. For remote work, build a strong Kaggle profile and GitHub portfolio. ML engineering and data science are among the highest-paid roles you can get from Pakistan.

07

Hardware & Embedded Programming Roadmap

Programming the physical world. From blinking an LED to building real embedded systems with microcontrollers and FPGAs.

Phase 1

Electronics Basics & Arduino

Start with basic circuits: resistors, capacitors, LEDs, and breadboarding. Get an Arduino Uno (available at Daraz and local electronics shops in Lahore's Hall Road or Islamabad's Jinnah Super). Write your first programs in C to blink LEDs, read sensors, and control motors. Arduino's official tutorials and Paul McWhorter's YouTube series are great starting points.

Phase 2

C Programming & Microcontrollers

Get proper with C: pointers, memory management, bitwise operations, structs. Move from Arduino to bare-metal programming on STM32 or ESP32 boards. Learn to read datasheets, configure peripherals (GPIO, UART, SPI, I2C), and write interrupt handlers. Embedded Systems Programming on YouTube by Quantum Leaps is excellent.

Phase 3

RTOS & Communication Protocols

Learn FreeRTOS for multitasking on microcontrollers: tasks, semaphores, queues, and scheduling. Understand communication protocols in depth: UART, SPI, I2C, CAN bus, and MQTT for IoT. Build a project that communicates wirelessly using ESP32 WiFi/BLE. The Mastering STM32 book by Carmine Noviello is solid for STM32-specific learning.

Phase 4

PCB Design & FPGAs

Learn KiCad (free) for PCB design: schematic capture, layout, and getting boards manufactured (JLCPCB ships to Pakistan cheaply). If you're into digital design, pick up an FPGA board and learn Verilog or VHDL. Nandland has the best beginner FPGA tutorials. This is where CS meets electrical engineering.

Phase 5

Linux, Drivers & Industry

Learn embedded Linux on a Raspberry Pi or BeagleBone. Write basic device drivers, understand the Linux kernel boot process. For careers: look at companies like NetSol, TPL Maps, automotive firms, and defense contractors in Pakistan. Internationally, embedded roles at Qualcomm, ARM, and automotive companies hire remote engineers. Embedded pays well because fewer people know it.

08

Robotics Roadmap

Where software meets the real world. Control systems, perception, planning, and building robots that actually do things.

Phase 1

Foundations: Math, Physics & Programming

Robotics needs linear algebra (transformations, rotations), calculus (for control theory), and basic physics (kinematics, dynamics). Get comfortable with Python and C++. If you haven't already, build something with Arduino to understand actuators and sensors hands-on. MIT OpenCourseWare has free courses on all the math you need.

Phase 2

ROS & Simulation

Learn ROS 2 (Robot Operating System), the standard framework for robotics software. Install it on Ubuntu, learn topics, services, actions, and launch files. Use Gazebo for simulation so you can test robot behaviors without hardware. The official ROS 2 tutorials and The Construct (online robotics academy) are your go-to resources.

Phase 3

Perception & Computer Vision

Robots need to see. Learn OpenCV for image processing, then move to 3D perception with depth cameras and LiDAR. Understand SLAM (simultaneous localization and mapping) and point cloud processing. Cyrill Stachniss's YouTube lectures on SLAM are fantastic. For deep learning-based perception, use YOLO for object detection and integrate it into ROS.

Phase 4

Control Systems & Motion Planning

Learn PID control, then move to state-space methods and model predictive control (MPC). For motion planning: A*, RRT, and potential fields. Brian Douglas's Control Systems YouTube channel makes controls intuitive. Build a line-following robot, then a robot arm that picks and places objects. NUST and GIKI robotics labs have strong communities around this.

Phase 5

Competitions & Research

Join your university's robotics team or start one. Compete in NERC (National Engineering Robotics Contest), or international challenges. For research, look into reinforcement learning for robotics, sim-to-real transfer, and manipulation. Companies like Waymo, Boston Dynamics, and drone startups hire robotics engineers. Pakistan has growing drone and agricultural robotics sectors too.

09

Generative AI Roadmap

LLMs, image generation, RAG, fine-tuning, and building AI-powered products. The fastest-growing field in tech right now.

Phase 1

Understand the Landscape

Before building, understand what exists. Play with ChatGPT, Claude, Midjourney, and Stable Diffusion. Understand the difference between LLMs, diffusion models, and multimodal models. Watch Andrej Karpathy's "Intro to Large Language Models" talk. Read the Attention Is All You Need paper (or at least a visual explanation of it). Know what transformers are and why they changed everything.

Phase 2

Building with APIs

Start building apps using the OpenAI API, Anthropic API, or Google Gemini API. Learn prompt engineering: few-shot prompting, chain-of-thought, system prompts. Build a chatbot, a document summarizer, or a code assistant. Use LangChain or LlamaIndex for chaining LLM calls. This is where you can start shipping real products fast.

Phase 3

RAG & Vector Databases

Retrieval-Augmented Generation is how you make LLMs work with your own data without fine-tuning. Learn embeddings, vector databases (Pinecone, ChromaDB, Weaviate), and chunking strategies. Build a "chat with your PDF" app or a knowledge base assistant. This is what most real GenAI products actually use in production.

Phase 4

Fine-Tuning & Open Source Models

Learn to fine-tune open models like Llama, Mistral, or Qwen using Hugging Face Transformers and LoRA/QLoRA. Understand when fine-tuning makes sense versus RAG versus prompt engineering. Use Google Colab or Lambda Labs for GPU access. For image generation, learn Stable Diffusion and ComfyUI. The Hugging Face course is free and practical.

Phase 5

AI Agents & Production

Build AI agents that can use tools, browse the web, and complete multi-step tasks. Learn about function calling, agent frameworks (CrewAI, AutoGen), and evaluation. For production: learn to deploy models, handle latency, manage costs, and build feedback loops. GenAI engineer roles are popping up at Pakistani startups and remote positions. Ship projects, write about them, and share on LinkedIn. This field rewards builders.

10

MLOps & DevOps Roadmap

The infrastructure side. How code gets from a laptop to production, and how ML models get deployed and monitored at scale.

Phase 1

Linux & Scripting

Get comfortable on the Linux command line. Learn file navigation, permissions, processes, and package management. Write Bash scripts for automation. Set up an Ubuntu VM or WSL on Windows. Learn Git properly: branching, merging, rebasing, and PR workflows. Linux Journey (linuxjourney.com) and The Missing Semester by MIT are great free resources.

Phase 2

Containers & Docker

Learn Docker: images, containers, Dockerfiles, volumes, and networking. Containerize a web app and a Python ML project. Learn Docker Compose for multi-container setups. This is the single most important DevOps skill. Everything in production runs in containers. Docker's official Getting Started guide and TechWorld with Nana on YouTube are solid resources.

Phase 3

CI/CD & Cloud

Set up CI/CD pipelines with GitHub Actions (easiest to start with) or GitLab CI. Automate testing, building, and deployment. Learn one cloud platform: AWS (most jobs), GCP, or Azure. Start with EC2, S3, and RDS on AWS. AWS Free Tier gives you 12 months of free usage. The roadmap.sh DevOps roadmap is a great visual reference.

Phase 4

Kubernetes & Infrastructure as Code

Learn Kubernetes for container orchestration: pods, deployments, services, and ingress. Use Minikube or Kind locally. Learn Terraform for infrastructure as code and Ansible for configuration management. This is where you go from managing one server to managing hundreds. KodeKloud has excellent hands-on Kubernetes labs.

Phase 5

MLOps & Monitoring

For MLOps specifically: learn MLflow for experiment tracking, DVC for data versioning, and model serving with BentoML or TensorFlow Serving. Set up monitoring with Prometheus and Grafana. Learn about model drift and retraining pipelines. DevOps/MLOps roles are well-paid in Pakistan and great for remote work. Companies like Afiniti, Teradata, and consulting firms like Deloitte hire from Pakistan.

11

System Administration Roadmap

Keeping servers alive, networks running, and systems secure. The backbone of every tech organization.

Phase 1

Linux Administration

Install and configure Ubuntu Server or Rocky Linux. Learn user management, file permissions, systemd services, cron jobs, and package management (apt, dnf). Set up a home server or a VPS on DigitalOcean ($5/month or free student credits). Linux Administration Handbook and RHCSA certification prep material on YouTube are great resources.

Phase 2

Networking Fundamentals

Understand the OSI model, TCP/IP, DNS, DHCP, firewalls (iptables/nftables), and SSH. Configure a web server with Nginx or Apache. Set up reverse proxies and SSL certificates with Let's Encrypt. Learn subnetting and basic network troubleshooting (ping, traceroute, netstat, tcpdump). Professor Messer's CompTIA Network+ videos are free and comprehensive.

Phase 3

Security & Hardening

Learn system hardening: disable unnecessary services, configure firewalls, set up fail2ban, manage SSH keys, and keep systems patched. Understand SELinux/AppArmor basics. Learn about common vulnerabilities and how to protect against them. For certifications, CompTIA Security+ is a good starting point. Security skills are increasingly important as Pakistani companies digitize.

Phase 4

Automation & Scripting

Automate everything repetitive with Bash scripts and Python. Learn configuration management with Ansible: write playbooks to configure servers, deploy applications, and manage users at scale. Use cron for scheduled tasks and learn log management with rsyslog or the ELK stack (Elasticsearch, Logstash, Kibana). Good sysadmins automate themselves out of repetitive work.

Phase 5

Cloud & Career

Move from bare metal to cloud: learn AWS (EC2, VPC, IAM, CloudWatch) or Azure. Get certified: AWS Solutions Architect or RHCE carry weight in Pakistan and internationally. Sysadmin roles at Pakistani banks (HBL, UBL), telecom companies (Jazz, Telenor), and enterprises (Systems Limited, TPS) pay well and are stable. Remote sysadmin/SRE roles are also common.

Hundreds of students prep together here. Share tips, ask questions, get past papers.

Join WhatsApp Group → Share Resources →