One match a week, no endless swiping, and an algorithm trying to figure out what U of T students actually want. The Row sat down with the students behind Yuni to talk campus dating, AI matchmaking, oddly specific preferences, and whether Toronto students need to get off the apps and actually go on a date.
At U of T, romance has some serious competition. There are midterms every other week, club meetings squeezed between lectures, case competitions, labs, networking events and the ongoing academic sport of convincing yourself that next week will somehow be less busy.
Dating can very easily end up somewhere near the bottom of the to-do list.
That is exactly the problem the students behind Yuni are trying to solve. Built around university dating, Yuni takes a noticeably different approach from the endless swipe, match, message, ghost cycle that has come to define dating apps. There is no mobile app to download and no feed of hundreds of profiles to work through. Instead, Yuni operates through text and iMessage and sends users a curated match each week.
The concept is almost aggressively simple: one carefully selected match, once a week, with the goal of getting two people off their phones and onto an actual date.
The Row sat down with Boris Tishkov, a Rotman Commerce student and Yuni’s CEO, and Ediz Cagan Uysal, a U of T Computer Science student and Yuni co-founder who works on the platform’s development and matching system, to figure out whether an algorithm can actually fix U of T’s dating problem.
The Problem Might Not Be Dating Apps. It Might Be U of T.
Before Yuni became a major part of their lives, Boris describes his university experience as a lot of studying, business classes and case competitions. For Ediz, studying Computer Science did not exactly create endless amounts of free time either. There were assignments, projects, exams and the pressure to keep grades up. Dating had to compete with everything else demanding their attention.
That experience helped inspire one of the central ideas behind Yuni: university students might want to date, but they do not necessarily want to spend hours trying to date.
When we asked what makes dating at U of T particularly difficult, their answer was straightforward.
Time.
Students are busy. Academics come first for a lot of people, and dating gets pushed further down the priority list.
Then there are the expectations. According to Boris and Ediz, U of T students also tend to have pretty high standards. You are surrounded by ambitious, intelligent and highly involved people, so it is not exactly surprising that students might expect the same things from a partner.
The result is a dating pool of people looking for someone smart, attractive, interesting, ambitious and compatible while simultaneously having approximately three free hours between now and finals.
Not exactly ideal conditions for romance.
One Gross Tinder Message and a Better Idea
The idea for Yuni was not only born from being busy. Boris recalled a female friend receiving what he described as a “really disgusting message” on Tinder. It raised a bigger question for the group.
Why had dating online become a process where users spend huge amounts of time swiping through strangers, maintaining multiple conversations and dealing with unwanted messages before they even make it to a real date?
The team wondered whether they could take most of that process away. Instead of giving someone hundreds of options and hoping they eventually find a good one, what if a system could choose one person with a strong chance of mutual interest?
As Boris explained during our interview, the idea was essentially to have AI find you “one really good person, one time per week,” leaving you to spend your time actually meeting people instead of endlessly sorting through them online.
No hours of swiping.
No graveyard of unanswered “hey” messages from three months ago.
Ideally, just a date.
So How Does Yuni Actually Work?
Yuni is technically not an app in the traditional sense. There is nothing to download. The service operates primarily through text and iMessage. Users create a profile and tell Yuni what they are looking for, including their preferences, interests and dealbreakers. Preferences are submitted by end of day on Tuesday.
Then comes what Yuni calls The Wednesday Drop. Every Wednesday at 7 p.m., users receive a personalized match. Before deciding whether they are interested, users can see information including their match’s first name, photos and shared interests. The more private information used to create the match, including a person’s specific preferences and dealbreakers, is not shown to the other user.
If the interest is mutual, the pair finds a time that works and Yuni helps curate the date at a public location near campus. The company describes the dates as casual and low-pressure, generally lasting around one to two hours.
Which is significantly less commitment than spending three weeks exchanging “how was your weekend?” messages on Hinge.
Okay, But How Does an Algorithm Pick Your Date?
This is where things get considerably nerdier. During our interview, Boris and Ediz explained that the goal is not simply to find someone who checks boxes on a profile. They are trying to predict the likelihood that both people will actually say yes to each other.
Yuni’s published privacy policy gives an even more detailed look behind the curtain. The matching pipeline begins with what Yuni calls hard filters. These include factors such as age range, sexual-orientation compatibility, geography and any dealbreakers supplied by a user. After that, the system calculates a compatibility score based on stated interests, lifestyle and what Yuni describes as “vibe” answers.
Then AI enters the equation.
Large language models are used to create additional tags from profiles and photos, giving the system more information to work with when evaluating compatibility. Another stage uses an LLM to score how a potential first interaction between two people might go. All of that produces a ranked recommendation.
But the algorithm does not necessarily get the final word. Yuni’s own policy states that a founder can review the recommendation, adjust it or choose a different pairing.
In other words, the system is not only asking: Does this person technically meet your requirements? It is attempting to answer a much harder question: Are these two people actually likely to want to meet each other?
Ediz explained that the system can also learn from previous decisions. If someone repeatedly turns down a certain type of profile, for example, that behaviour can become another signal for the matching process.
The ultimate goal is not to build an algorithm that can magically identify your soulmate. It is to find someone you might actually want to meet.
Your “Type” Might Be Less Important Than You Think
Possibly the most interesting thing Yuni has learned so far has nothing to do with building an app. It is that people may not actually know their own type as well as they think they do.
Users can give Yuni a pretty detailed picture of what they want. The platform collects information about interests, lifestyle, relationship intentions, physical characteristics and personal preferences alongside dealbreakers and compatibility requirements.
And students can get very specific. During our conversation, Boris and Ediz described users looking for particular heights, programs, sports, ethnicities, religions, appearances and combinations of characteristics.
The problem is that those stated preferences do not always predict what happens in real life.
The founders say they have already seen people start dating Yuni matches who did not perfectly fit the criteria those users originally submitted. Someone might insist that their match needs to be a particular height, study a specific subject or participate in a certain activity, only to meet someone outside that exact description and discover they actually click.
The founders connected that observation to research examining the gap between the qualities people say they want in a partner and the people they are attracted to once they actually interact face-to-face.
It is an interesting contradiction for a matchmaking service built around collecting preferences. Sometimes the best match might not be the person who perfectly resembles the checklist you wrote beforehand.
The algorithm needs to get two mutually interested people into the same room.
After that, chemistry gets a vote.
What Are U of T Students Actually Looking For?
Naturally, we had to ask.
From the patterns Boris and Ediz have observed so far, physical attraction appears to play one of the largest roles in whether someone initially accepts or rejects a match. Personality and common interests matter too, but the founders said appearance frequently becomes one of the strongest initial factors.
Religion is another recurring preference. For students who consider religion an important part of their lives, sharing a faith can become an important compatibility requirement.
Yuni’s own privacy policy confirms that the platform can collect particularly detailed matching preferences, including age ranges, dealbreakers and optional partner-ethnicity preferences, though those private preferences are not shown to other users.
And sometimes what users request is so specific that even a fairly large university dating pool cannot easily satisfy it.
The founders said one of the hardest parts of building the system is simply capturing what people actually mean when they describe their type. A detailed explanation of someone’s personality, lifestyle and values gives an algorithm quite a bit to work with.
“Tall, handsome, brown hair” does not.
At the time of our interview, the team said it was working on a newer system intended to capture those preferences more accurately.
We Made Them Stereotype Your Faculty
All this data about attraction obviously meant we had to use our time responsibly. So we asked Boris and Ediz to assign dating stereotypes to four classic U of T categories.
According to the highly peer-reviewed Yuni School of Dating Research:
Rotman: Avoidant, probably wearing business attire and potentially a little arrogant.
Engineering: Extremely locked in, always studying and possibly so busy that dating has not even entered the equation.
Life Sci: More of a “nerdy type,” interested in learning about different things.
Humanities: More extroverted, more likely to enjoy partying and trying different activities.
We will be accepting formal complaints from all four faculties at a later date.
What Happens When an AI Matchmaker Has to Think About Safety?
Of course, any platform encouraging strangers to meet has a much less entertaining problem to solve.
Safety.
During our interview, Boris and Ediz repeatedly emphasized that they want Yuni to feel connected to the university community rather than like meeting a completely anonymous person from the internet.
The platform is 18+ and currently centres on supported university communities. Yuni can require users to verify their university email and phone number as part of the process.
Its policies also say that uploaded photos are automatically checked for content and quality, including nudity, minors, obviously fraudulent or generated images and whether the main profile photo contains a visible face.
Once two users actually decide to meet, Yuni directs dates toward public locations near/on campus. Users can also report someone or request that a particular person not be matched with them again.
Yuni’s community guidelines prohibit harassment, threats, stalking, discriminatory or hateful content, unwanted sexual behaviour and several other forms of abuse. The company says it aims to review reports of abusive behaviour within 24 hours.
Building a Dating Service While Attending U of T Is About as Relaxing as It Sounds
There is a little irony here. Yuni exists partly because university students do not have enough time. Its founders are now university students running a startup that takes up essentially all of theirs.
For Boris, a typical day can include marketing, filming content, talking to users, organizing events and figuring out what needs to be improved next. Immediately after our interview, he was heading out to get coffee with Yuni users.
Ediz’s days live more on the technical side. There are messaging systems to maintain, unusual edge cases to fix, matching results to review and scheduling problems that need to work even when two students have timetables apparently designed by mortal enemies.
He also works on improving the matching system from week to week based on what users say they want and how satisfied they are with the matches they receive.
Even with an algorithm doing the actual calculations, somebody still needs to make sure the system works. Every week brings another list of things to fix, improve, test or organize.
Startup life, apparently, is not helping either of them escape the U of T workload problem.
Building Yuni Has Changed How They Think About Dating Too
Working this closely with hundreds of students’ dating preferences has also changed how the founders themselves think about relationships.
For Boris and Ediz, one of the biggest lessons has been watching how often real relationships ignore the rules people originally set for themselves.
People come in thinking they know exactly what they want. Then they meet someone. Suddenly, the exact program, exact height or exact checklist matters a little less.
That is part of why the founders seem much more interested in getting students onto dates than keeping them interacting with Yuni itself.
Success is not somebody spending an hour inside the product. Success is somebody leaving it to meet another person.
Their Solution to U of T Dating Might Not Actually Be More Technology
Toward the end of our conversation, we asked the founders to finish one sentence: The U of T dating scene would improve immediately if ______.
Their answer? There were more events.
Not another algorithm. Not better profiles. Not some new premium dating feature.
Events.
For a campus with tens of thousands of students, it can still be bizarrely difficult to meet somebody organically. If your daily schedule consists of lecture, Robarts, lecture, TTC, repeat, the love of your life could be sitting four tables away and you would probably never speak.
Boris argued that the usual options can also become repetitive. There are club events, mixers, pubs and nights out, but not always many campus events specifically designed to make meeting someone new feel easy or interesting.
So Yuni has started moving offline too.
On August 25, Yuni hosted a U of T version of the Pop the Balloon dating-show format at IDK Social Bar. It went viral with millions of views across Canada and the rest of the world. The format had participants meet potential dates while holding balloons. If they lost interest, they would pop the balloon.
Yuni’s version let students attend as audience members or apply to participate, with a Yuni profile required for entry.
On September 6th, Yuni is hosting a frosh matchmaking party.
Then on September 10, Yuni is hosting a Red Flags party with a live matchmaking dating show at NEST nightclub with over 300 people. Wear a red shirt and you’ll be given tape to write a red flag on it at the party. The party is 19+, with doors opening at 10 p.m., and Yuni-profile holders are being offered discounted admission of up to 40%.
For a company centred around AI matchmaking, Yuni’s vision for improving campus dating is surprisingly analogue.
Get people together. Make something interesting happen. Give them a reason to talk.
More Than 100 Dates Later
Yuni’s ambitions are not exactly modest. During our interview, Boris and Ediz told The Row that the platform had already facilitated more than 100 dates on campus at UofT.
The site also currently reports an average date rating of 4 out of 5, says 93 per cent of respondents want a second date, and shows more than 1,100 students joined.
Those are Yuni’s own reported numbers rather than independently verified statistics, but they suggest the experiment has moved beyond a handful of friends trying out a student startup.
And the founders do not seem particularly interested in judging success by downloads anyway. There is not even an app to download.
The metric they kept returning to during our conversation was much simpler:
Dates. People actually meeting each other.
They want the average U of T student to go on more dates than they did before. Boris even predicted that Yuni could lead to “ten times” as many dates happening around campus as the service continues to grow.
Ambitious?
Absolutely.
But when your starting hypothesis is that U of T students barely have time to date at all, there is plenty of room to work with.
So, Could Yuni Produce a Marriage?
Obviously, we asked.
The response was basically immediate.
Yes.
The logic is pretty simple. If enough people meet, some of them will start dating. If enough of those relationships last, statistically, eventually somebody is probably going to have to explain at their wedding that an AI built by U of T students introduced them.
We desperately hope they save the original Wednesday Drop.
What Would the Matchmakers Pick for Themselves?
After spending all this time thinking about everyone else’s love lives, we had one final important question. What does the person building your dating algorithm actually consider a good date?
Boris’s answer was refreshingly uncomplicated. Grab boba, walk around campus and talk. For him, the point of the first date is getting to know the other person. He said he would rather use the time to build an actual connection and figure out whether they are compatible than anything else.
Ediz gave a similar answer: coffee, conversation, finding common interests and seeing where things go from there.
Which, after a conversation about algorithms, AI, preferences, matching probabilities, lifestyle signals and behavioural patterns, leaves Yuni with a surprisingly human conclusion.
You can collect every preference imaginable.
You can calculate compatibility.
You can optimize the match.
You can tell an algorithm your ideal height, program, personality, interests, religion and hair colour.
But eventually, the algorithm has to get out of the way.
Somebody still has to sit across from somebody else.
Somebody still has to start the conversation.
Somebody still has to figure out whether there is actually a spark.
And at U of T, finding the time might still be the hardest part.