001Campus●○○
Course Schedule Agent
Builds a conflict-free timetable from degree rules, professor ratings, and live seat availability, then grabs slots the second they open.
Signal: Students already pay for seat-sniping bots every registration day.
// A free, fully online hackathon for student teams. Pick a validated idea, prove it like a founder, then ship a working AI agent. The best builders don't just win, they leave with a startup on the launchpad and a fast-track into our internship.
Most hackathons end with a demo nobody remembers. This one is built differently: you spend week one thinking like a founder and week two building like an engineer. Form a group, claim an idea, publish your analysis, ship the agent. If it's good, we help you take it further than a weekend.
Find 2 to 5 students who argue well and ship fast. Mix skills: someone who codes, someone who talks to users, someone who can't stop asking "but who pays for this?". Register your team below. Solo entries are allowed, but founding is a team sport.
Claim one of the 100 validated ideas below, or bring your own valid use case. Then create a public git repo and submit your detailed analysis: the problem, who has it, evidence the pain is real, who pays, existing alternatives, your agent architecture, and a scoped MVP plan. The analysis repo is due Saturday, June 20.
Build the solution in the same repo. It must be genuinely agentic: an AI that perceives context, makes decisions, and takes actions toward a goal, not a thin chatbot wrapper. Commit as you go; your history is part of the story. Rough edges are fine, fake demos are not.
Final submission: working agent, README with setup steps, and a short demo video (under 5 minutes) pitched like you're raising your first round. Winners announced the following week, with written feedback for every team that finishes.
// 100 agentic AI ideas, each anchored to a real pain and a real buyer, so you spend week one validating, not wandering. Filter, search, or hit "deal me one" and let fate choose. Claiming an idea drops it straight into your registration. Duplicates across teams are allowed: execution is the moat.
Builds a conflict-free timetable from degree rules, professor ratings, and live seat availability, then grabs slots the second they open.
Signal: Students already pay for seat-sniping bots every registration day.
Scans thousands of scholarships, matches them to a student's profile, tracks deadlines, and drafts first-pass applications.
Signal: Billions in scholarship money goes unclaimed every year because nobody can find it.
Watches job boards, tailors your resume per posting, submits, and runs polite follow-ups until you get an answer.
Signal: The average student sends 100+ applications per season, mostly copy-paste drudgery.
Splits work, tracks who actually did what from commits and docs, nags slackers, and compiles the final submission.
Signal: Every student has horror stories; professors increasingly ask for contribution reports.
Records lectures, builds structured notes, then generates spaced-repetition decks and quizzes timed to your exam calendar.
Signal: Note-taking apps and flashcard apps both thrive; nobody bridges them automatically.
Ingests the syllabus and past papers, predicts likely topics, and builds an adaptive day-by-day study plan that reshuffles when you slip.
Signal: Coaching institutes charge heavily for exactly this planning service.
Reads lost-item posts across campus groups and chats, matches them to found reports by description and location, and brokers the handoff.
Signal: Every campus has scattered "lost my ID card" posts that die unseen.
Learns your interests, finds fests, talks, and club events worth your time, registers you, and blocks your calendar with travel buffers.
Signal: Campus events routinely under-fill while students complain there's nothing to do.
Interviews students conversationally about sleep, noise, cleanliness, and guests, then proposes compatible matches with an honest compatibility report.
Signal: Bad roommate matches drive dropouts and transfers; housing offices match by spreadsheet.
Collects food complaints and ratings, finds patterns, forecasts demand per dish, and drafts weekly menu changes for the vendor.
Signal: Mess committees run on angry group chats; vendors waste food cooking blind.
Watches receivables, sends escalating but polite payment reminders, reconciles incoming payments, and flags accounts that need a human call.
Signal: Late payments are a top-3 killer of small businesses worldwide.
Monitors Google, Yelp, and food-delivery reviews, drafts on-brand replies, and escalates genuine service failures to the owner with context.
Signal: Response rate measurably affects local rankings; most owners reply to nothing.
Tracks sales velocity per SKU, predicts stockouts, drafts purchase orders at the right moment, and learns seasonal patterns.
Signal: Small retailers still reorder on gut feel and lose sales to empty shelves.
Keeps hours, menus, photos, and contact info in sync across Google, maps apps, and directories, and fixes drift automatically.
Signal: Wrong opening hours online is one of the most common local-business complaints.
Confirms appointments, reschedules conflicts conversationally, and backfills cancellations from a live waitlist for salons, clinics, and tutors.
Signal: No-shows cost appointment businesses 10-20% of revenue.
Turns a plumber's site photos and voice notes into an itemized, priced, professional quote sent before they're back in the van.
Signal: Trades lose jobs to whoever quotes first; many quote days late or never.
Reads what sold this week, drafts a content calendar with posts, captions, and offers, and queues everything for approval in one tap.
Signal: Owners know they should post; the median shop account is silent for months.
Chats with applicants for shift jobs, checks availability and basic fit, answers their questions, and books interviews straight into the manager's calendar.
Signal: Most hourly applicants never even get a reply; managers drown in no-fit resumes.
Tracks licenses, permits, tax filings, and renewal deadlines for a business's exact location and type, preps the paperwork, and nags until it's filed.
Signal: Small businesses pay real fines every year for deadlines they simply forgot.
Spots regulars who stopped showing up, figures out their pattern, and crafts a personal nudge or offer that doesn't feel like spam.
Signal: Retaining a customer is famously cheaper than finding one; small shops track neither.
Manages reminders, tracks adherence, reorders refills before they run out, and alerts a caregiver when doses are repeatedly missed.
Signal: Half of chronic-condition patients don't take meds as prescribed; the cost is enormous.
Reads medical bills and explanation-of-benefits documents, spots errors and denials worth fighting, and drafts the appeal letters.
Signal: A large share of denied claims are overturned when appealed; almost nobody appeals.
Triages symptoms conversationally, finds the right kind of doctor nearby with open slots, books the visit, and preps a question list.
Signal: People bounce between wrong specialists for months; booking friction delays care.
Watches rehab exercises through the phone camera, corrects form in real time, logs progress, and adapts the plan between clinic visits.
Signal: Most physio patients abandon home exercises within two weeks; recovery stalls.
Makes a friendly daily call or chat with an elderly relative, notices changes in mood, speech, or routine, and escalates anomalies to family.
Signal: Adult children living far away pay for any reliable signal that mom is okay.
Turns meal photos into nutrition logs, spots patterns against your goals, and delivers a weekly nutritionist-style review with concrete swaps.
Signal: Manual calorie logging has massive churn; photos remove the friction.
Answers a small clinic's phone and chat after closing: books and reschedules appointments, answers FAQs, and routes emergencies correctly.
Signal: Clinics miss dozens of booking calls a day; every missed call is lost revenue.
Translates blood work into plain language, tracks your values across years, and generates the right questions to ask at the next appointment.
Signal: Patients google cryptic lab acronyms within minutes of every report.
Predicts class demand for gyms and studios, fills empty slots from waitlists, and proposes schedule changes that match real attendance.
Signal: Studios run half-empty classes at peak rent; scheduling is pure guesswork.
Tracks sleep, screen time, and activity signals the user opts into, predicts when a habit streak is about to break, and intervenes early with the right nudge.
Signal: Habit apps celebrate streaks but do nothing at the moment they're about to die.
Finds every recurring charge across your statements, flags the ones you stopped using, and walks through cancellation or negotiation for you.
Signal: The average person underestimates their subscription spend by hundreds per year.
Turns a shoebox of receipts and a travel calendar into filed, policy-compliant expense reports with zero manual data entry.
Signal: Employees procrastinate expense reports for months, leaving their own money on the table.
Tracks gig income across platforms, sets aside the right tax amount automatically, estimates quarterly dues, and preps filing checklists.
Signal: First-year freelancers routinely get blindsided by tax bills they never planned for.
Compiles comparable listings, vacancy data, and your payment history into a data-backed negotiation letter and a renewal strategy.
Signal: Most tenants accept the first renewal number; landlords expect pushback and price for it.
Reads bank and UPI statements, categorizes everything, finds the leaks (fees, duplicates, lifestyle creep), and automates savings rules you approve.
Signal: Budgeting apps show charts; almost none take action on what they find.
Audits telecom, utility, and hospital bills against your plan and usage, then runs the dispute process: letters, portal forms, escalations.
Signal: Billing errors are common and profitable; disputing them is designed to exhaust you.
Explains what's happening in a beginner's portfolio in plain language, simulates "what if" moves safely, and quizzes understanding before any real action.
Signal: A generation is entering markets via apps with zero guidance and high regret.
Reads loan and credit agreements before you sign, flags risky clauses, prepayment traps, and junk fees, and compares against market norms.
Signal: Nobody reads the fine print; predatory clauses count on exactly that.
Runs a shared budget for couples or flatmates, settles "who owes what" automatically, and referees overspend conversations with neutral data.
Signal: Money fights are a top cause of household conflict; splitting apps only track, never mediate.
Logs every purchase from receipts and emails, remembers warranty windows, and files return or claim paperwork before deadlines pass.
Signal: Consumers forfeit billions in valid warranty claims because nobody tracks the dates.
Adjusts reading passages to a child's exact level, asks Socratic questions, and quietly raises difficulty as comprehension improves.
Signal: Reading-level gaps in one classroom can span four grades; one teacher can't bridge them alone.
Applies the teacher's rubric to a stack of essays or assignments, drafts per-student feedback, and surfaces the cases needing human judgment.
Signal: Teachers spend 5+ hours a week grading; feedback quality drops with fatigue.
Stages real-life conversations (job interview, market haggling, doctor visit) in a target language, remembers your recurring mistakes, and drills them.
Signal: Language app users plateau at vocabulary; speaking practice is the universal gap.
Compiles each student's week (grades, participation, flags) into a short, kind digest for parents, in their language, with suggested home support.
Signal: Parents hear from school twice a year; problems surface months too late.
Assesses what you can actually do, diffs it against a target job's real requirements, and builds a project-based path that closes the gap.
Signal: Course platforms sell content; learners actually want a route to a paycheck.
Guides students through problems Socratically without ever giving the answer, and reports genuine effort and stuck-points to the teacher.
Signal: Schools are banning answer-bots; a tutor teachers can trust is the open seat.
Turns any coaching center's curriculum into question banks, mock tests, and per-student weak-spot drills, formatted and print-ready.
Signal: Tutoring centers pay content writers per worksheet; demand never stops.
Builds a realistic shortlist from grades and budget, tracks every deadline and document, and gives structured essay feedback without writing it.
Signal: Families pay consultants thousands; most students navigate it with spreadsheets and panic.
Gives students with learning disabilities live captions, simplified parallel summaries, and structured notes tuned to their specific needs.
Signal: Accessibility offices are legally required and chronically under-resourced.
Pairs students with alumni by goals and personality, brokers the intro, suggests agendas, and keeps the mentorship alive with gentle structure.
Signal: Alumni want to help and students want guidance; the match never happens by itself.
Routes pull requests to the right reviewer, drafts a first-pass review of the boring parts, and nags until the queue is empty.
Signal: Review latency is a top developer complaint in every engineering survey.
Watches CI history, identifies flaky tests statistically, quarantines them, bisects to the introducing commit, and files evidence-rich tickets.
Signal: Teams silently retry red builds daily; flakes erode trust in the whole suite.
Reconstructs an incident timeline from logs, alerts, and chat, drafts the blameless postmortem, and tracks whether action items actually ship.
Signal: Postmortems arrive weeks late or never; action items die in docs.
Opens upgrade PRs with changelog summaries, runs the test suite, attempts auto-fixes for breaking changes, and explains what it couldn't fix.
Signal: Dependabot opens PRs; nobody merges them because nobody trusts them.
Diffs API docs and READMEs against the actual code on every merge, flags drift, and opens fix PRs with corrected examples.
Signal: "The docs are wrong" is the most common onboarding complaint in software.
Answers a new hire's "where does X happen" and "why is it like this" questions with linked code, history, and the PRs that explain decisions.
Signal: New engineers take months to ramp; senior time is the bottleneck.
Hunts idle instances, unattached volumes, and oversized databases, builds an evidence file per finding, and proposes terminations for approval.
Signal: A third of cloud spend is waste; finance sees the bill, engineers see nothing.
Clusters production errors into root-cause groups, dedupes against existing tickets, and files new ones with reproduction context attached.
Signal: Error trackers fire thousands of alerts; triage is a dreaded rotation everywhere.
Detects new or changed UI strings on each merge, drafts translations in context, and opens PRs so every language ships together.
Signal: Non-English users wait weeks for features because translation is an afterthought.
Scans every build for WCAG violations, ranks them by user impact, and opens scoped fix PRs with before/after screenshots.
Signal: Accessibility lawsuits keep rising; most teams have zero a11y tooling in CI.
Reads smart meter data, finds the appliances and habits burning money, and takes approved actions like schedule shifts and thermostat tweaks.
Signal: Energy prices keep climbing; utility dashboards inform but never act.
Connects restaurant and bakery surplus to shelters and community fridges, handling timing, quantities, and pickup logistics automatically.
Signal: A third of food is wasted while food banks run short; coordination is the only gap.
Diagnoses a broken appliance from photos and symptoms, estimates repair cost versus replacement, and books a vetted local repairer.
Signal: Right-to-repair laws are spreading; people default to landfill out of pure uncertainty.
Answers "which bin does this go in" instantly, tracks dorm-level recycling rates, and runs competitions that actually change behavior.
Signal: Contamination ruins entire recycling batches; campuses have sustainability budgets to spend.
Estimates the footprint of each purchase from statements, finds your top three impact levers, and suggests specific, realistic swaps.
Signal: People want to act on climate but get generic advice instead of personal numbers.
Aggregates neighbor interest, models shared savings, gathers quotes, and shepherds the group through paperwork to installation.
Signal: Group solar is dramatically cheaper, yet almost nobody can organize their street.
Watches meter and usage patterns for homes and societies, detects leaks early, alerts residents, and dispatches a plumber with the evidence.
Signal: A silent leak can waste thousands of liters before the bill exposes it.
Intercepts your shopping intent, checks used marketplaces for the same item in good condition, and handles the haggling and pickup logistics.
Signal: Resale is booming, but searching five marketplaces per purchase is too much work.
Pools dead electronics across a neighborhood, schedules certified recycler pickups when volume justifies the trip, and tracks what got diverted.
Signal: Everyone has a drawer of dead phones; single-household pickups don't pay for themselves.
Builds a planting calendar from your exact location and climate data, then runs care reminders, pest alerts, and harvest timing all season.
Signal: Gardening surged post-pandemic; beginners kill plants from timing, not neglect.
Coordinates parcel deliveries for a gated community or hostel: one drop point, smart notifications, and routed last-meter handoffs.
Signal: Guards manage parcel chaos with paper registers; deliveries fail daily.
Plans an entire house move: collects mover quotes, builds the packing checklist, schedules utility transfers, and updates your address everywhere.
Signal: Moving is a universally hated multi-week project with zero coordination tools.
Tracks mileage, fault codes, and service history for small van or truck fleets, predicts failures, and books service in slack windows.
Signal: One surprise breakdown can erase a small operator's weekly margin.
Tracks every shipment you send or receive, detects loss and damage, and files compensation claims with carriers before windows close.
Signal: Carriers owe refunds for late deliveries that almost no one claims.
Matches commuters at one campus or office by route and schedule, handles trust and cost-splitting, and re-plans when someone bails.
Signal: Parking is over capacity everywhere; carpool boards die from coordination overhead.
Analyzes pick patterns for small warehouses, recommends where each SKU should live, and re-slots continuously as demand shifts.
Signal: Pickers walk kilometers a day for layouts nobody has revisited in years.
Pools daily ingredient orders across street vendors in one area to unlock wholesale prices, then coordinates a single split delivery.
Signal: Vendors each pay retail and lose an hour daily fetching the same supplies.
Monitors your flights and trains, predicts disruption early, and rebooks within your rules before the queue at the help desk even forms.
Signal: Travelers discover cancellations at the gate; rebooking speed decides who gets the last seat.
Watches temperature sensors on perishable shipments, alerts on excursions, and triggers contingency actions: reroute, re-ice, or claim.
Signal: Vaccine and food spoilage in transit is a billion-dollar gap that small shippers feel hardest.
Tracks return windows across all your online orders, batches drop-offs into one optimal trip, and preps every label and QR code.
Signal: People miss return windows constantly; returns are the most hated chore of online shopping.
Turns long videos and streams into platform-native shorts: finds the hooks, cuts, captions, and schedules each clip where it performs best.
Signal: Creators pay human clippers per video; short-form is where discovery happens now.
Matches a creator's audience to brands actually sponsoring in their niche, drafts the pitch with real stats, and tracks negotiations to a signed deal.
Signal: Mid-size creators leave money on the table because outreach is a second job.
Triages comments across platforms, replies in the creator's voice to the easy ones, escalates the sensitive ones, and mines them for content ideas.
Signal: Engagement drives the algorithm, but answering comments doesn't scale past 10k followers.
Mines a newsletter archive and turns each issue into threads, scripts, carousels, and eventually a structured digital product.
Signal: Writers sit on years of content and republish none of it.
Takes a raw recording through edit notes, chapters, show notes, titles, quote graphics, and platform uploads, end to end.
Signal: Post-production is why most podcasts die before episode ten.
Handles a creator's DM flood: answers FAQs in their voice, spots superfans and business inquiries, and surfaces only what needs the human.
Signal: Creators report DM guilt as a real burnout factor; real opportunities drown in noise.
Generates merch concepts from a creator's catchphrases and art, polls fans, collects pre-orders, and only triggers production when demand clears cost.
Signal: Unsold merch boxes are a rite of passage; pre-order validation flips the risk.
Scans platforms for stolen reuploads of a creator's work, gathers evidence, files takedowns, and tracks each case to resolution.
Signal: Freebooting steals billions of views; manual takedowns take hours per incident.
Generates thumbnail and title variants, rotates them live, monitors click-through, and locks in winners automatically per video.
Signal: Top channels employ full-time thumbnail artists; CTR is the whole game.
Turns a creator's existing videos and posts into a structured course: outline, lesson scripts, slides, quizzes, and a launch email sequence.
Signal: Courses are the highest-margin creator product, and the production lift is why most never ship.
Tracks crop prices across nearby markets in real time, factors transport cost, and tells a farmer exactly where and when to sell today's harvest.
Signal: Price spreads between markets an hour apart can swallow a farmer's entire margin.
Diagnoses crop disease from a leaf photo, prescribes treatment with exact dosage, and connects to the nearest supplier who stocks it.
Signal: Crop disease destroys up to 40% of yield globally; agronomists are scarce in rural areas.
Combines weather forecasts, crop stage, and soil data into simple daily watering instructions delivered over SMS or WhatsApp.
Signal: Both over- and under-watering cut yields; the data exists, the last mile doesn't.
Lets a corner-shop owner manage stock, reorders, and customer credit ledgers entirely by voice, in their own language, while serving customers.
Signal: Tens of millions of micro-retailers run on memory and paper notebooks.
Matches farmers and small businesses to government schemes they actually qualify for, fills the applications, and tracks them through approval.
Signal: Scheme money lapses unspent every year because applying is a maze.
Turns an artisan's product photos into polished listings with pricing guidance, posts them across marketplaces, and handles inquiries in any language.
Signal: Skilled makers earn a fraction of retail because middlemen own the digital shelf.
Tracks each animal's feed, weight, and symptoms via quick voice notes, spots early illness patterns, and escalates to a vet with the full history.
Signal: One missed illness can spread through a herd; record-keeping is nonexistent.
Manages shared ownership of expensive equipment (drills, tillers, projectors): bookings, deposits, condition checks, and fair-use scheduling.
Signal: A drill is used 13 minutes in its lifetime; whole neighborhoods buy one each.
Coordinates seasonal farm work: matches workers to harvests by skill and location, schedules crews, and handles fair, instant payouts.
Signal: Crops rot in fields while willing workers sit idle two villages away.
Picks a food truck's daily location and menu using local events, weather, and past sales, then announces it to nearby regulars automatically.
Signal: Location choice swings a truck's daily revenue 3x, and most pick by habit.
Did you find a real pain and prove it? Week one's analysis repo carries this score.
Does your AI perceive, decide, and act in a loop? Tool use, memory, and judgment beat prompt wrappers.
Does it run, end to end, on the real problem? An honest rough demo beats a polished fake one.
Could this become a company? Clear buyer, credible path to revenue, and a team that could carry it.
We're not handing out hoodies. We're looking for people who can take an idea from nothing to running in production, because that's exactly what we do all day.
Winning team members skip straight to the take-home stage of our AI Engineer Intern hiring process. Your hackathon repo is your application; no CV screening, no first-round call.
Three months of mentorship from the KartiLine team to take your idea toward real users: positioning, pricing, architecture, and the unglamorous parts nobody blogs about.
A public repo with a founder-grade analysis and a working agent, plus written feedback from our team. That's a portfolio piece that outlasts any certificate, and winners get featured on kartiline.com.
// Register your team before kickoff on Saturday, June 13, 2026. The git repo link can come later; the team can't.
Any student team of 2 to 5 members, from any college, anywhere. Fully online, completely free. Solo entries are allowed, but founding is a team sport: find your people.
No. The vault exists so you don't burn week one hunting for a problem. Bring your own idea for any valid use case, as long as the solution is genuinely agentic: an AI that perceives, decides, and acts toward a goal. Multiple teams can claim the same idea; execution is the moat.
Your system should run a loop: read real context, make a decision, take an action through tools or APIs, observe the result, and continue. A chat UI over a prompt is not agentic. An agent that watches invoices, decides which are overdue, drafts reminders, and escalates the stubborn ones, that is.
Two checkpoints, one public git repo. By Saturday, June 20: your detailed idea analysis (problem, evidence, who pays, alternatives, agent architecture, MVP scope). By Saturday, June 27: the working implementation, a README with setup steps, and a demo video under 5 minutes.
Anything you can legally access: any LLM, any framework, any stack. Free tiers are enough to win. We judge the problem, the agent's design, and the working result, not the size of your API bill.
The KartiLine Agentic AI Hackathon is a free two-week online hackathon for student teams of 2 to 5, running Saturday, June 13, 2026 to Saturday, June 27, 2026. Teams choose one of 100 validated agentic AI ideas or bring their own valid use case, submit a public git repository with a detailed founder-style analysis by June 20, then implement a working AI agent by June 27. Judging: problem insight 25%, agentic depth 25%, working demo 25%, venture potential 25%. Winners fast-track into the KartiLine AI Engineer Internship and receive three months of founder mentorship. Register at https://kartiline.com/hackathon.html#register.