EXLAC.COM — PERSONALISED CAREER ROADMAP

Fahmida's 1-Year AI Agent Builder Roadmap

Twelve months. Twelve real-life projects. One clear path from C beginner to AI Agent developer — building skills companies will pay for in 2026 and beyond.

Candidate: Fahmida Alif Duration: 12 months Track: AI Agent Builder 2–3 hrs/day, evenings

Fahmida, this is your plan

Your assessment showed something clear: you are drawn to AI, you have unusually high patience (5/5), and you are willing to work independently. You wrote that you want to "solve real-life problems" with AI — that is exactly what this roadmap is built around.

Every month for the next year, you will build one real, useful project — not toy exercises. By month 12 you will have a public GitHub portfolio with twelve real projects, the technical foundation of an AI Agent developer, and the kind of work that wins remote international contracts.

You said your main worry was "keeping up with the rapidly changing technology landscape." Good news: the fundamentals in this roadmap (Python, APIs, databases, prompts, agents) do not change. Tools change; principles stay. By learning the principles deeply, you will never be left behind. প্রতি মাসে একটি বাস্তব project — এক বছরে ১২টি project, GitHub-এ সাজানো, যা দিয়ে remote work-এর জন্য client বা company-কে দেখাতে পারবে।

Candidate Profile (From Your Questionnaire)

Programme
B.Sc. CSE, Southeast University
Year
First Year (current)
Daily Time
2–3 hours, evenings
Patience
5/5 — very high
Self-Direction
4/5
English
Good
4-Year Goal
Remote work, international clients
Current Skills
C basics, AI tools (ChatGPT)

Career Direction Scores

AI Agent Development
100%
AI / Machine Learning
84%
Data Science / Analytics
84%
Mobile App Development
63%
Web Development
55%

Programme Structure

Four quarters, twelve months, twelve real-life projects — each building on the last

Quarter 1 — Months 1-3
Programming Foundations
Python, OOP, files, web basics, Git, GitHub.
Quarter 2 — Months 4-6
APIs, Databases, & AI
FastAPI, SQL, the OpenAI/Claude API, your first AI features.
Quarter 3 — Months 7-9
RAG, UIs, & Bots
Vector search, React basics, Telegram bots.
Quarter 4 — Months 10-12
Real Agents & Capstone
WhatsApp bots, tool-using AI agents, capstone project.
Monthly format: every month has (1) Learning topics, (2) a Real-Life Project — something genuinely useful, not a toy exercise, and (3) notes on how it maps to your university coursework so you stay ahead in class too.
Q1
M1-3

Quarter 1 — Programming Foundations

Move from C to Python, learn Git, and ship your first three real-life projects.

Month 1 Python Foundations
From C to Python — the language of AI. Your first programs will run in the first week.
L What You Learn
  • Python syntax (very similar to C, but cleaner): variables, types, conditionals, loops
  • Functions, lists, dictionaries, string manipulation
  • Reading and writing files
  • Using pip and your first Python virtual environment
  • How to ask AI tools (ChatGPT / Claude) to explain code you don't understand
P Real-Life Project
Personal Expense Tracker (CLI) You Will Actually Use This
A command-line tool that lets you record daily expenses (food, transport, books), saves them to a file, and shows a monthly summary. Build it for yourself — once it works, use it every day.
  • Add expense with amount, category, date, note
  • View all expenses, filter by category or date
  • See total spent this month and breakdown by category
  • Save everything to a local expenses.json file
PythonJSONCLI
Deliverable: Public GitHub repository with README + a screen recording (Loom) showing the tool in use over a few real expense entries.
U University Connection
  • Python's syntax will make your C++ course (your hardest subject) feel easier — the concepts are the same, only the punctuation differs.
  • The list and dictionary operations here will preview your Data Structures course (your favourite!).
Month 2 Object-Oriented Python & Git
Organise your code properly and learn to version control everything you build.
L What You Learn
  • Classes, objects, methods in Python (OOP — the same OOP you study at university)
  • Inheritance, encapsulation, simple design patterns
  • Modules and packages — how to split a big program into files
  • Git fundamentals: init, add, commit, push
  • GitHub: creating repositories, README files, public profile
P Real-Life Project
Study Schedule Organiser For Your Own Semester
A Python application using OOP that helps you plan your week — classes, assignments, exam dates, study hours. Use it to actually plan your semester. Push every commit to GitHub.
  • Classes for Course, Assignment, StudySession
  • Add upcoming deadlines and get reminders of what's due this week
  • Track how many hours you studied for each course
  • Generate a weekly study plan based on deadlines
  • Saved to JSON; restored on next run
Python OOPJSONGitGitHub
Deliverable: GitHub repository with clean OOP design + at least 10 commits + a README explaining how to use it.
U University Connection
  • The OOP concepts you practise here will appear in your university OOP and Software Engineering courses next semester.
  • Push your university assignments to GitHub too — build your public profile from now.
Month 3 Web Basics — HTML, CSS, JavaScript
Learn how the web works so you can show your work to anyone, anywhere.
L What You Learn
  • HTML: structure, headings, lists, forms, links, images
  • CSS: colours, fonts, layout (Flexbox & basic Grid)
  • JavaScript basics: variables, functions, DOM manipulation
  • Responsive design — making pages look good on mobile
  • Free deployment via GitHub Pages or Netlify
P Real-Life Project
Your Personal Portfolio Website Public — Build Your Brand
A responsive personal portfolio website featuring your name, photo, short bio, university, and your first two projects from Month 1 & 2. Deploy it free on GitHub Pages at fahmida-alif.github.io.
  • Clean home page with your name, headline, and contact info
  • About section — who you are, where you study, what you're building
  • Projects section — cards linking to your GitHub repositories
  • Mobile-responsive (looks good on phone)
  • Live on a real public URL
HTMLCSSJavaScriptGitHub Pages
Deliverable: Live URL of your portfolio + GitHub repository. This URL goes on your CV from now on.
U University Connection
  • Your university web courses (later semesters) will go deeper into this — you'll already be ahead.
  • This portfolio will hold your final-year project, internship records, and any awards.
Q2
M4-6

Quarter 2 — APIs, Databases & First AI

Build proper backends, store data in databases, and write your first AI-powered code.

Month 4 FastAPI — Building Real Backends
Move from CLI tools to real web services that any app can talk to.
L What You Learn
  • What an API is and what a backend service does
  • FastAPI fundamentals: routes, request validation with pydantic
  • HTTP methods (GET POST PUT DELETE) and status codes
  • Returning JSON, handling errors
  • API documentation auto-generated via FastAPI's /docs
P Real-Life Project
Bookmark Manager API Replace Browser Bookmarks
A FastAPI service where you can save any web link with a title, tags, and notes. Add, list, search, and delete bookmarks via simple API calls. Save it to a JSON file for now.
  • POST /bookmarks — save a new link with title, URL, tags, notes
  • GET /bookmarks — list all (with optional tag filter)
  • GET /bookmarks/search?q=... — search by keyword
  • DELETE /bookmarks/{id} — remove a bookmark
  • Auto-generated documentation at /docs
FastAPIPydanticPython
Deliverable: GitHub repo + screenshot of the /docs page + a Postman collection demonstrating each endpoint.
U University Connection
  • Networks course concepts (HTTP, request-response) come to life when you build a real API.
  • Use this exact project as a reference when your Software Engineering course teaches REST.
Month 5 SQL & Databases
Stop saving to JSON. Use a real database the way every production system does.
L What You Learn
  • SQL essentials: CREATE TABLE, SELECT, INSERT, UPDATE, DELETE
  • JOIN, GROUP BY, ORDER BY
  • SQLite (file-based, perfect for projects) and how to connect from Python
  • Designing simple tables and relationships
  • Avoiding SQL injection with parameterised queries
P Real-Life Project
Small Shop Inventory System Solve a Real Local Problem
A FastAPI + SQLite inventory system for a small local shop (family member, neighbour, or imaginary). Track products, stock levels, and sales. Show the shop owner a real demo — that is your proof of value.
  • Add products with name, price, current stock
  • Record sales (deduct stock automatically)
  • List all products, low-stock products (alert when below threshold)
  • Daily and weekly sales reports
  • SQLite database, restored every time the app starts
FastAPISQLiteSQL
Deliverable: GitHub repository + database schema diagram + a Loom (5 min) demonstrating it managing real or realistic sample data for a small shop.
U University Connection
  • Your Database Systems course (likely in semester 4 or 5) will already feel familiar.
  • This is exactly the kind of project that turns into a paid freelance gig later in your degree.
Month 6 Your First AI Code — LLM APIs
From here, every project gets smarter. This is the month AI clicks for you.
L What You Learn
  • What an LLM is (without the heavy math) — how it predicts, completes, reasons
  • The OpenAI API and the Anthropic Claude API — calling them from Python
  • Prompt engineering basics: roles, instructions, examples, constraints
  • JSON / structured outputs — getting reliable machine-readable answers
  • Cost control, rate limits, and free tiers
P Real-Life Project
AI Study Assistant For Your Own Coursework
A FastAPI app where you paste a topic from your university course (a concept, theorem, code snippet, exam question) and the AI explains it in simple language with examples. Use it for your own studying every week.
  • POST /explain — paste any topic, get a clean explanation
  • Three modes: "Explain like I'm a beginner," "Quiz me on this," "Give me 3 worked examples"
  • Saves your study history (SQLite from last month)
  • Returns answers in structured JSON
  • Simple web page (from Month 3 skills) to use it from the browser
FastAPIOpenAI/Claude APISQLiteHTML
Deliverable: GitHub repository + a Loom showing it explaining 3 real topics from your current syllabus, plus a short note on how much it helped you.
U University Connection
  • This tool will help you study every other course you take from this month onward.
  • If your university teaches an AI / Intelligent Systems course later, you will already be well ahead.
Q3
M7-9

Quarter 3 — RAG, UIs & Bots

Make AI answer questions about your documents, build proper UIs, and deploy your first messaging bot.

Month 7 RAG — AI That Knows Your Documents
Teach AI about your own notes, PDFs, and course material. This is the technique behind almost every real AI product.
L What You Learn
  • What embeddings are: turning text into numbers AI can compare
  • What a vector database is (use ChromaDB — free and Python-native)
  • The RAG pattern: retrieve → augment → generate
  • Chunking documents, indexing, top-k similarity search
  • Grounded answers with citations — "answer only from this context"
P Real-Life Project
PDF Q&A Bot for Your Course Notes Real Exam Prep Tool
An app that lets you upload your university PDF notes, indexes them, and answers your questions from inside the notes — with citations to the exact pages. Use it during exam revision.
  • Upload multiple PDF files (e.g. all your DSA lecture notes)
  • Automatic chunking and embedding
  • Ask a question, get an answer + source pages
  • Polite refusal when answer is not in your notes
  • Web interface for upload and chat
FastAPIChromaDBOpenAI/ClaudePyPDF
Deliverable: GitHub repository + a Loom showing it answering three real questions from your course notes with correct citations + one polite refusal.
U University Connection
  • Use this tool for every major exam from now on — your grades will likely improve.
  • This is the foundation of every modern AI chatbot — you are learning a top-paid 2026 skill.
Month 8 React Basics & Modern UI
Add a proper interface to your AI projects so anyone can use them — not just developers.
L What You Learn
  • What React is and why most modern apps use it
  • Components, props, state with useState
  • useEffect and fetching data from your FastAPI backend
  • Tailwind CSS for fast, clean styling
  • Building from a Vite template, deploying to Vercel free tier
P Real-Life Project
AI Recipe Generator (Web App) Useful Every Day
A nice-looking web app where you type the ingredients you have at home, and the AI suggests three recipes you can make right now. You will actually use this. Deploy publicly on Vercel.
  • React frontend with a clean ingredient-input UI
  • Calls a FastAPI backend, which calls the LLM
  • Returns 3 recipes with steps, time, difficulty
  • "Save favourite" feature stored locally
  • Mobile-responsive, deployed at a real URL
ReactTailwindFastAPIOpenAI/ClaudeVercel
Deliverable: Live public URL + GitHub repository + a Loom demonstrating it generating recipes from your actual kitchen ingredients.
U University Connection
  • React is now the industry-standard frontend — way ahead of what most universities teach.
  • This project alone qualifies you for entry-level frontend freelance work.
Month 9 Your First Telegram Bot
Move your AI out of the browser and into people's pockets — on Telegram.
L What You Learn
  • How to create a bot with @BotFather
  • The Telegram Bot API basics
  • Python python-telegram-bot library
  • Handling text messages, commands, and inline buttons
  • Renting a tiny VPS (~$5/month) and deploying the bot 24/7
P Real-Life Project
Personal AI Assistant on Telegram You Will Chat With This Daily
Your personal AI bot on Telegram. Ask it questions, set reminders, summarise long messages, translate Bangla to English. Live on a real VPS, available to you anywhere.
  • /start, /help, and free-text Q&A
  • /remind 1h Drink water — reminds you in 1 hour
  • /translate — Bangla to English or vice versa
  • /summary — summarise any long message
  • Deployed on a VPS, running 24/7
Telegram Bot APIPythonVPSOpenAI/Claude
Deliverable: Bot username + GitHub repo + a Loom of you chatting with your live bot.
U University Connection
  • You now know how to deploy a service on a server — an essential professional skill many graduates lack.
  • This bot will solve real problems for you — reminders, translations, quick Q&A — for years.
Q4
M10-12

Quarter 4 — Real Agents & Capstone

Build agents that take actions in the real world, integrate WhatsApp, and finish the year with a portfolio-grade capstone.

Month 10 WhatsApp Bot Integration
The most-used messaging app in Bangladesh — learning to build for it opens real local revenue.
L What You Learn
  • WhatsApp integration options — Baileys (free, unofficial) vs the official Cloud API
  • QR-code pairing for a dedicated number
  • Receiving and sending messages, media, buttons
  • Session persistence across restarts
  • Bridging Telegram (Month 9) and WhatsApp to a single agent backend
P Real-Life Project
Small-Shop Order Bot on WhatsApp For a Real Local Shop
A WhatsApp bot that takes orders for a small local shop (the same shop from Month 5, or a new one). Customers send "I want 2 kg sugar" and the bot confirms, records the order, and notifies the shopkeeper. This is real local business automation.
  • Customer sends a natural-language order — the bot parses it
  • Checks inventory (from your Month 5 shop database)
  • Confirms the order, calculates total, asks for delivery address
  • Notifies the shopkeeper of new orders
  • Polite handling when an item is out of stock
WhatsAppBaileysFastAPISQLiteOpenAI/Claude
Deliverable: GitHub repo + Loom (5 min) showing real or realistic orders being placed and confirmed via WhatsApp.
U University Connection
  • This is a real product worth selling. Many local shops in Dhaka would pay for this.
  • An ideal CV item for any internship application.
Month 11 Tool-Using AI Agents
Until now, your AI answered questions. This month, it takes actions.
L What You Learn
  • Agent loops: plan, act, observe, iterate
  • Tool use / function calling: giving the AI a list of actions it can take
  • Designing safe, deterministic tools (clear inputs, predictable outputs)
  • Memory: short-term conversation, long-term per-user storage
  • Loop safety — maximum iterations, timeouts, error handling
P Real-Life Project
Smart Customer Service Agent Sellable Product
A multi-tool customer service agent for a real or fictional small business. Tools: search_faq, check_order_status, create_support_ticket, escalate_to_human. The agent decides which tool to use for each customer message. Reachable from your portfolio website, Telegram, and WhatsApp.
  • 4 different tools the agent can call
  • Conversational memory per customer
  • Hard cap of 5 reasoning steps per customer message
  • Friendly escalation when it cannot help
  • Available on all three channels (web + Telegram + WhatsApp)
LangGraphFastAPISQLiteMulti-channel
Deliverable: GitHub repo + a Loom (5 min) demonstrating the agent handling four distinct customer situations correctly, using different tools.
U University Connection
  • You now build the kind of system real companies pay engineers to build — while still a 2nd-year student.
  • This is a strong showcase for SUST/IUT/BUET-level competitive programming & project showcases too.
Month 12 Capstone — Your Signature Project
One year, twelve projects. This month, you build the one that defines what you can do.
L What You Learn
  • How to scope a real product end-to-end (problem → users → design → build → ship)
  • Polishing a project for portfolio — README, demo video, architecture diagram
  • Writing about your work clearly — LinkedIn posts, project blog
  • Reviewing the year and planning Year 2 of your CSE degree
P Real-Life Project
Multi-Channel AI Agent Platform Capstone
Combine everything: pick a real problem you care about (tutoring, healthcare bookings, women's safety information, freelance lead capture, anything that matters to you), and build a complete AI agent platform around it. Web UI + Telegram + WhatsApp + your own knowledge base + tools.
  • One unified backend (Month 4-11 combined)
  • Three channels: website, Telegram, WhatsApp
  • RAG over your domain knowledge (Month 7)
  • Multiple tools the agent can call (Month 11)
  • Deployed on a real VPS at a real domain
  • A polished README, demo Loom, and architecture diagram
PythonFastAPIReactOpenAI/ClaudeChromaDBTelegramWhatsAppVPS
Deliverable: Live URL + GitHub + Loom + LinkedIn post announcing the project. By end of Month 12, your portfolio has 12 real projects.
U University Connection
  • You are now ahead of most 3rd and 4th-year CSE students in real-world skill.
  • Your GitHub is filled with one year of consistent, real work — rare and valuable.
  • You are ready for paid internships, remote freelance work, or international scholarship applications.

Tips for the Year Ahead

Habits that quietly compound over twelve months

Consistency Beats Intensity

Two focused hours every evening will beat a ten-hour weekend marathon. Show up daily, even if some days you only do thirty minutes.

One Project a Month is the Rule

Even if it is small, ship it. A finished imperfect project is worth ten unfinished perfect ones. By month twelve, you will have a portfolio that most graduating students will not have.

Always Push to GitHub

Every single project, every single week. Your green-square contribution graph is your proof of consistency. International recruiters and remote clients look at this first.

Use AI as Your Always-Available Tutor

When you are stuck, ask ChatGPT or Claude to explain. But never paste code you don't understand — always read line by line and ask "why does this work?" until you can explain it to a friend.

About "Keeping Up With Changing Tech"

Your assessment said this worries you. Here is the truth: the basics (Python, APIs, databases, prompts, agents) have been stable for years and will be for years more. Tools change — fundamentals don't. Master the fundamentals this year, and you will adapt to any new tool quickly.

Don't Wait Until You "Feel Ready"

Start Month 1 this week. You will never feel ready. The way to feel ready is to start, struggle, and keep going.