Find work
Filters · 1
Our heating and cooling business misses most calls after 6pm, and those are often emergencies. We want an AI receptionist that answers, works out if it is urgent, books a visit into our calendar or texts the on-call technician, and sends us a summary of every call. It must sound natural, spell names and addresses back, and never promise a price. You choose the platform; tell us the monthly running cost.
We need 50 original word problems that today's models get wrong at least some of the time. Secondary-school to first-year-university level: rates, probability, geometry, a little number theory. Each problem needs one correct answer, a full worked solution, and a note on the trap (the step a model tends to skip). Nothing copied from textbooks or the web; we check. We will run each problem through two models before accepting it, and pay for every problem at least one of them fails.
We want our support team to search past tickets by meaning, not keywords: "customer can't log in after password reset" should find every similar ticket even when the words differ. Our stack is Postgres and TypeScript. Add pgvector, embed the 5,000 tickets, keep new ones embedded, and build a small search page with filters for date and product. Show us how good it is on 30 test searches we provide.
Half of our patients message us on WhatsApp, many in Spanish. We want a bot that books, moves and cancels appointments in our booking system, in either language, and passes anything medical to the front desk. Our booking system has an API and we have a WhatsApp Business account. You will build the flow, connect both, and train our front desk for 30 minutes on how to take over a conversation.
Our refund bot is supposed to follow a strict policy: no refunds after 30 days, never promise money it cannot give, never reveal another customer's order. We want you to break it. Find 40 distinct prompts or conversations that make it break a rule. For each, write what you tried, what it said, and which rule it broke. Variety counts: 40 versions of the same trick count as one. You get a test account and the written policy. Please do not test anything outside the bot.