Side Projects That Land Internships: 14 Project Ideas By Major

A junior at Georgia Tech I coached last spring sent the same resume to 41 companies. Three callbacks. Brutal. Then she shipped a Chrome extension that auto-tagged her tabs by class. 280 lines of JavaScript, a 90-second Loom demo, a public GitHub repo. And dropped the link into her resume header. Next batch of 22 applications? Seven interviews, three offers, one of them at a fintech she'd been told she had "no shot" at. Same student. Same GPA. The only variable was a working artifact a recruiter could click. That's the entire pitch for side projects for resume students: a clickable thing beats a polished bullet point every time. Recruiters scan resumes in 6 to 8 seconds and they're looking for one signal. can this person actually build. A GitHub link with green commits answers that before they finish your name. This guide walks through 14 specific projects you can actually build between now and your next application cycle, broken out by major. CS folks get the biggest list because they have the most options, but I've included mechanical engineering, business, and design alternatives that punch above their weight. No vague suggestions like "build a website." Every project below has scope, a stack, and a story you can tell in an interview. The trick isn't picking the fanciest idea. It's picking one you'll actually finish, deploy somewhere with a public URL, and write a 200-word README about. A finished tetris clone with a live demo will out-perform a half-built distributed system every single time. Promise.

Why side projects for resume students matter more in 2026

Recruiters are drowning. The Simplify x Pitt CSC Summer 2026 internships repo on GitHub crossed 50,000 stars this year, which is the polite way of saying every CS student in the world is applying to the same 800 roles. AI screening tools filter resumes by keyword density before a human ever sees them, and when a human does see yours, they have maybe 10 seconds to decide. A live link to a working project is the cheat code. It signals three things at once: you can ship, you can write, and you care enough to do it without a grade attached. The Hacker News "Ask HN: What side projects landed you a job" threads are stuffed with stories of folks who got cold-emailed by recruiters because of one repo with 40 stars. That's the bar. Not 5,000 stars, not a YC-grade startup. Forty.

The other reason this matters more now: AI tools mean coursework no longer differentiates anybody. Everyone's CS 101 PSET looks the same when half the class used Claude. What can't be faked is a deployed app with real users, a Streamlit dashboard with a public URL, a Chrome extension installed by 30 people, a Kaggle notebook with 500 views. Those leave a fingerprint. When a recruiter clicks your GitHub and sees five repos with READMEs, screenshots, and live demos. and recent commits. You've crossed a threshold most applicants haven't. Honest take: at this point, side projects aren't optional for landing first internships. They're table stakes.

CS Project #1: A Chrome extension that solves your own problem

Best first project. Period. Chrome extensions are tiny (manifest.json + popup.html + content.js gets you 80% there), they live in a real users' browsers, and the Chrome Web Store gives you a public URL recruiters can install. Pick something annoying you do daily. Tab management. Highlighting Stack Overflow answers. Auto-filling job application forms. Counting how many minutes you spend on Reddit before the panic-study session.

Stack: vanilla JavaScript, the Chrome Extensions API (Manifest V3), maybe one fetch call to a free API. Time budget: a weekend for v1, another weekend to polish. Publish it on the Chrome Web Store. The $5 one-time developer fee is the cheapest credential money can buy. Record a 60-second screen capture, embed it in your README, and put the install link in your resume. A friend who graduated UMich in 2024 shipped a "Did I already apply here?" extension that highlighted previously-applied job postings on LinkedIn. 1,800 installs by graduation. Two internship offers explicitly mentioned it in the interview.

CS Project #2: A financial dashboard with real market data

Finance dashboards photograph well. Open the repo, see a chart with candlesticks and moving averages. Recruiter brain lights up. Use Python with the `yfinance` library (it's free, hits Yahoo Finance, and works in five lines of code), pull stock data for 5 to 10 tickers, then build a Streamlit or Plotly Dash front-end with selectable date ranges, indicators (50-day EMA, RSI, MACD, Bollinger Bands), and a comparison view.

Deploy to Streamlit Community Cloud (free) or Render (free tier covers this easily). Don't make it ugly. Spend 90 minutes on the design. Add a "watchlist" feature where users save tickers. Bonus: pull headlines from a news API and tag them as sentiment-positive or negative using a free Hugging Face model. The github.com/DhruvAthaide/StockMarketFinancialDashboard repo is a decent reference for scope; you want roughly that size. This is one of the strongest portfolio projects internship recruiters at fintech firms (Stripe, Plaid, Robinhood, smaller props) actually click on.

CS Project #3: A Kaggle dataset analysis with a story

Kaggle has 50,000+ public datasets. Pick one that interests you personally. Spotify streaming data, NBA shooting logs, FIFA player stats, Airbnb listings in your city. And write a notebook that answers a non-obvious question. Not "what's the average?" but "do songs with explicit lyrics get more skips before 1:30, and does that vary by genre?" The framing matters more than the model.

Use pandas, seaborn, maybe scikit-learn for one prediction. Publish the notebook publicly on Kaggle and cross-post on Medium or your own blog with the charts embedded. When your notebook hits ~500 views and 20 upvotes, Kaggle gives you a "Notebooks Expert" tier badge that's a real credential. Worth every penny. And it's free. The data science Kaggle community is one of the most-checked places by data internship recruiters at IBM, Meta, and smaller analytics shops.

CS Project #4: An ML model deployed via Streamlit or Gradio

A model in a Jupyter notebook? Cute. A model with a public URL where strangers can upload an image and get a prediction? Different league. Train something small. A CNN that classifies plant leaves as healthy/diseased, a sentiment classifier on Yelp reviews, a recommender on the MovieLens dataset. And deploy it to Hugging Face Spaces (free) using Gradio. The whole deployment file is maybe 30 lines.

This is the single project category that's exploded for 2026 internships because generative AI hiring is everywhere. Recommendation systems power Netflix, Spotify, every e-commerce store, and basically every FAANG product. A live recommender with a real-world dataset and a working Gradio frontend is a stronger portfolio piece than three static notebooks combined. Add a "model card" in your README explaining what the model can't do. That one section signals more maturity than the model itself.

CS Project #5: A Tetris clone (or any small game)

Sounds basic. Isn't. A working Tetris clone in vanilla JavaScript, Pygame, or with a real game engine like Godot shows you can manage state, handle input, run a game loop, and ship something playable. The Tetris-clone GitHub topic has hundreds of repos and the best-starred ones get 1k+ stars from people who just want to play it on lunch break.

Host it on GitHub Pages (free), make it playable in-browser, add a leaderboard with localStorage, then add Firebase for a real cloud-synced leaderboard. Now you've got a full-stack project camouflaged as a game. Bonus storyline for interviews: "I rewrote the collision detection three times before it stopped flickering" beats any rehearsed STAR-method answer. Skip it only if you genuinely don't like games. Your enthusiasm will leak through either way.

CS Project #6: A web scraper that does something useful

Scraping is a real engineering skill. Rate limits, retries, parsing weird HTML, headless browsers. And it's a fantastic side project because the output is concrete. Build something that scrapes a site you use weekly: campus housing listings, your school's class registration page, a sneaker drop site, concert tickets on AXS, internship postings. Output to a Google Sheet, an SQLite database, or trigger a Discord/email alert when something changes.

Use Python's `requests` + `BeautifulSoup` for static sites, Playwright or Selenium for JS-heavy ones. Respect robots.txt. That's both an ethics thing and an "I know what I'm doing" interview signal. The Israel Dryer Web-Scraping-Projects repo is a solid model for portfolio packaging. Pair the scraper with a tiny Flask or FastAPI endpoint so others can query your data, and you've quietly built a backend service.

CS Project #7: A REST API that other projects can use

Recruiters love seeing a REST API because it tests architecture thinking, not just hacking. Build something narrow but useful: a college course difficulty API, a meal-plan calorie API, a "is this dorm haunted" API (joke project, real engineering). Use FastAPI for Python or Express for Node, write tests with pytest or Jest, deploy to Railway/Render/Fly.io, generate OpenAPI docs.

The killer move: build two projects, where the second one consumes the first. Your scraper feeds your API, your API powers your dashboard. Suddenly you're not "a student with random projects". You're "a student who builds systems." That phrasing shows up word-for-word in performance reviews at every big tech company. Use it in your interview.

CS Project #8: A Discord bot that lives in a real server

Discord bots are the most underrated portfolio projects internship candidates can ship because they have users. Real ones. Build a bot for your dorm's Discord, your club's server, or a study group: track LeetCode streaks, post daily problems, schedule mock interviews, summarize lecture chats with the OpenAI API, run polls. The discord.py library has 16.1k GitHub stars and is well-documented; discord.js has 26.7k.

Deploy to Railway or Fly.io's free tier and add a `!stats` command that shows uptime and command count. When you get to your interview, you can say "47 people in my dorm use this every day". That's a real metric, not a vanity one. The CamdenClark/CSCHDiscordBot repo shows exactly how to scope a bot around a real club use case (resume reviews, mock interviews). Steal the structure shamelessly.

Mechanical Engineering: A CAD project with FEA + a build log

For ME students, the equivalent of a GitHub repo is a portfolio website with CAD renders, exploded views, and FEA simulation results. Pick one project: a hydraulic gripper arm, a small wind turbine blade optimized for low wind speeds, a quadcopter frame, or a 3D-printable bike component. Model it in Fusion 360 or SolidWorks (both have free student licenses), run a static stress analysis in ANSYS or the built-in simulation tool, document failure modes, then build it if budget allows.

Write a build log on a personal site or Notion page: design constraints, three iterations, what broke, what you'd change. Recruiters at Tesla, Boeing, SpaceX, and smaller mech shops want to see iteration, not just final renders. WashU's career center literally tells engineering students to include "your role, tools used, key deliverables, quantitative outcomes". Bold those four things in every project description. Quantitative outcomes especially. "Reduced part weight 28% while maintaining safety factor of 2.1" beats "designed a bracket."

Business / Marketing: A real campaign with a public dashboard

Business students often default to "I wrote a case study". Fine, but case studies don't have public links. Better: run an actual marketing experiment. Start a niche Instagram or TikTok account around something you genuinely care about (campus food reviews, study setups, broke-student finance tips), grow it for 90 days, then publish a Notion dashboard with weekly metrics, A/B tested captions, and what worked. Even 800 followers from zero is enough proof.

Or: pick a local small business (your favorite coffee shop, a friend's Etsy store), offer to run their Google Ads or Meta Ads campaign for $200 budget, document the funnel. Get a screenshot of the ROAS. That's a real portfolio piece marketing internship recruiters at Procter & Gamble, Unilever, smaller DTC brands will actually pause on. Pair it with a Google Ads or HubSpot certification (both free) and you've got a teachable, hireable story.

Design: A full case study with the messy middle included

Design portfolios live or die on case study depth. Pick one project. Could be a hypothetical redesign of your university's course-registration system, a real client gig from a campus org, or your own side product. And build a 6-section case study in Figma: problem framing, user research (even 5 user interviews counts), 3-4 wireframe iterations, the final prototype, what you'd change, and measurable outcomes. The Figma community has a Portfolio Case Study Builder plugin that scaffolds the whole thing in seconds.

Show the messy middle. Hiring managers at design shops can spot a fake portfolio in 30 seconds; what they want is the wrong direction you went, the user feedback that killed it, and how you pivoted. Depth beats breadth here. Three deep case studies will outperform eight shallow ones. Host on a personal site (Framer, Webflow, or even a free Notion site) so it's clickable from your resume, not a PDF attachment.

How to package side projects for resume students

Building it is half the job. Packaging is the other half, and most students butcher this. Each project on your resume needs four things: a one-line summary that says what it does and who uses it, a tech stack (3-5 keywords matched to the job description), a quantitative outcome (users, accuracy, latency, GitHub stars, anything numerical), and a clickable link. Not "available on request." Not a screenshot. A live URL. Repeat: the entire point of side projects for resume students is clickable proof.

Your GitHub README is your interview before the interview. Add a header image (a screen recording GIF or hero screenshot), a 200-word "what it does" section, a "tech stack" block with badges, an install/run section, and a "what I'd do next" paragraph. That last section signals self-awareness. Interviewers adore it. Pin your top six repos on your GitHub profile. Set a custom profile README with a tldr of who you are. This stuff takes 90 minutes total and outperforms most students' entire portfolio sites.

Conclusion

The students who land internships in 2026 won't be the ones with the longest resume bullet points. They'll be the ones with three clickable links a recruiter can open in a new tab and explore for 90 seconds. That's the whole game. Pick one project from the 14 above, give yourself a two-weekend deadline, deploy it somewhere with a real URL, and write a README a stranger could understand. Then put the link at the top of your resume. The side projects for resume students playbook is honestly that simple. Most students just never finish step one. Be the one who does.

Frequently Asked Questions

How many side projects should I have on my resume as a student? Three to five solid projects beat ten shallow ones every single time. Recruiters scanning your resume have 6-10 seconds. They'll click maybe one link. Make that one link impossible to ignore. The sweet spot for personal projects resume sections is three projects with depth: one signature piece (your strongest), one that shows breadth (different language or stack), and one quirky or personal one that gives the interviewer something to ask about. If you have eight projects, pin the best six on GitHub and only list three on the resume itself.

Do side projects matter more than GPA for internships? Depends on where you're applying. For FAANG and quant firms with explicit GPA cutoffs (typically 3.5+), GPA gets your foot in the door. For 90% of other tech internships. Startups, mid-size product companies, agencies, smaller fintechs. A strong portfolio with a working project will outweigh a 3.2 vs 3.5 difference. The data from Simplify's Summer 2026 internships repo and Hacker News hiring threads is consistent on this: hiring managers prefer candidates who shipped something over candidates who memorized something. Build projects to get hired isn't just a slogan; for non-elite-school students, it's the actual playbook.

What's the difference between a "good" side project and a "great" one for internships? A good side project works locally and has a README. A great one is deployed at a real URL, has at least one external user (even a friend counts if it's real), has commits over multiple weeks (not a single weekend dump), and tells a story. Recruiters can spot weekend-only projects instantly. They look like tutorial hell. The great ones show iteration: v1, then bug fixes, then a feature based on user feedback. That arc of commits is what signals you'll handle real engineering work.

Is it worth open-sourcing my project or keeping it private? Open source it. Always. Even if it's embarrassing. Private repos signal "I'm afraid of feedback," public repos signal "I welcome it." Recruiters at every level can read code. A clean, well-commented public repo is one of the strongest portfolio projects internship signals you can send. If you're worried about quality, add a `KNOWN_ISSUES.md` file listing what's broken. That's actually a positive signal, not a negative one. It shows engineering maturity.

Can side projects help non-CS majors land internships too? Yes, in fact arguably more, because non-CS majors are competing in fields where evidence of initiative is rarer. A mechanical engineering student with a CAD portfolio and a build log is a unicorn next to a peer with just a GPA. A business student running a real Google Ads campaign with screenshots of the dashboard out-performs a peer with case-study-only experience. The bar is lower outside CS, so even a moderately ambitious side project breaks you out of the resume pile. Don't sleep on this.

How long does it take to build a portfolio-worthy side project? Two to four weekends for a solid v1 if you're focused. The Chrome extension, Tetris clone, and Discord bot projects above are genuinely doable in a weekend each. Bigger projects like the financial dashboard or deployed ML model need 3-4 weekends including deployment, README, and demo recording. Don't sandbag. Most students drag projects out for months because they keep adding features. Ship v1 in two weekends, then iterate.

Should I use AI tools like Claude or Copilot when building side projects? Yes, use them. But understand the code you ship. Recruiters in 2026 fully expect candidates to use AI assistants. What they're filtering for is whether you actually understand the code, can extend it, can debug it when something breaks. Build with Copilot, then close it and try to rewrite a function from scratch. If you can't, you have a learning gap to fix. The interview will catch this either way; better to find it now.

What if my side project gets zero stars or users. Does it still count? Yes. Stars and users are nice but not required. A well-built project with zero stars and a great README is still strong proof. What matters is: does it work, is the code readable, did you ship to production, can you talk about the tradeoffs? Most interviews never reference the star count. They reference your decisions. A small project where you can articulately defend every architectural choice will out-perform a 500-star project you copy-pasted from a YouTube tutorial.