Test It Yourself
“Oh? Junhak’s kids?”
Professor Shin Dongjin broke into a broad smile the moment he saw Kim Siwon’s email.
It was admirable that they had managed to send materials that looked credible despite the tight preparation window.
A business plan file that was larger than expected, along with four separate attachments sorted by use case.
“Then I’d better read it carefully.”
So, with a fond sort of curiosity, Professor Shin opened the attachments at once.
He even found himself thinking, So this is why Professor Hwang Junhak is pushing them so hard!
“Oh, this is clean. They really did put in some effort.”
But as he read through it lightly, the business plan turned out to be surprisingly serious in its structure.
The separate file attached for the stray dog promotion app was also written in excellent, concrete detail.
“What is this? I thought they were undergraduates?”
A little startled, Professor Shin opened every attachment again and looked through them from the beginning.
Then he checked the student information Professor Hwang had passed along as well.
The results did not match their age at all.
Leaning back in his chair, Professor Shin let out an admiring sigh.
“Hah. So that’s why he wanted to push them.”
The attachments Kim Siwon had sent looked nothing like the work of ordinary undergraduates.
They looked like files from a startup that had already secured at least several tens of billions of won in investment.
“It wasn’t just because they were hard-working kids. He really recognized them because they were good.”
Nodding, Professor Shin read through the files once more.
There was practically nothing to pick apart.
It was bold of them to prepare as if they were assuming his help in advance, but if he really did lend a hand, it looked as though they could move forward with a credible business right away.
“Then I guess I don’t need to treat them like ordinary college students.”
After a long, silent stretch of thought, Professor Shin stood up and made a call somewhere at once.
“All of you come to my lab tomorrow. By 3:30 p.m.”
***
Everyone from the club, myself included, headed to Professor Shin Dongjin’s lab wearing white shirts and slacks.
Of course, we had a hard time with the ones who insisted that dressing like developers was a matter of confidence.
“This is my first time in the veterinary medicine building.”
“Same here. It’s really clean.”
“It doesn’t smell like animals. Right?”
“Is that important right now? Aren’t you all nervous?”
“Being too nervous isn’t good either.”
I was nervous too.
But since students had such good things to say about Professor Shin, I was able to calm myself at least somewhat.
He was the kind of professor who, so long as you didn’t go against the school rules, would try his best not to give you an F.
Knock, knock, knock.
“Come in.”
But the moment we opened the lab door, we froze a little at the atmosphere inside, which was different from what we had expected.
The graduate students were lined up in a row, and at the end of it stood Professor Shin, watching us with a serious expression.
“Good afternoon. I’m Kim Siwon from the Department of Computer Engineering.”
Starting with me, all of us greeted him politely.
It felt like something straight out of the military.
The atmosphere was far stiffer than it had been with Professor Hwang Junhak, and everyone was visibly tense.
“You’re the representative here, Kim Siwon?”
“Yes, sir.”
“Then don’t get jittery like the people beside you. Just explain things. I printed out all the PPT materials you sent.”
At that, I looked around and saw that the graduate students and the professor were all holding copies of our presentation materials.
Everyone even had pens in hand, ready to take notes.
I calmly opened my laptop and finished preparing the presentation.
“Good afternoon. I’m Kim Siwon, a third-year undergraduate student in the Department of Computer Engineering. I’ll now present the business plan for an AI system that detects canine lung cancer and pneumonia through X-rays.”
Even on the title slide, I could see people making notes.
“This graph shows the rising number of people raising companion animals. The average annual growth rate over the past ten years is XX percent, and over the past three years, the growth has accelerated even further to XX percent.”
I went through the presentation step by step, exactly as I had prepared it.
The rising number of people with pets, the current state of the pet industry, an explanation of the product’s mechanism, and then the scale of the market share we could capture in the future and our marketing direction.
This was not a presentation meant simply to trumpet the brilliance of our technology, as I had in my previous life. It was a presentation designed to persuade the audience effectively.
“Isn’t there a little too little about our actual product?”
Understandably, the club members who had worked so hard on the product were slightly disappointed.
By contrast, the eyes of my targets, the professor and the graduate students, were shining.
“It looks really convincing. Right?”
“It does.”
Everyone seemed to have quite a lot they wanted to say.
“That’s all. If you have any questions about the presentation, please let me know.”
Every graduate student and the professor in front of me raised a hand.
Then, when the professor raised his hand too, all the other students lowered theirs.
It was a moment that made the subtle hierarchy immediately visible.
Professor Shin Dongjin. He really wasn’t an easy man.
“By what did you decide to make an AI that diagnoses only lung cancer and pneumonia? If you were including pneumonia, you could have added cardiomegaly and other diseases too.”
“In the case of humans, lung cancer is often detected through X-rays by looking for tumors, so I added it first. Pneumonia was added because it is one of the common conditions suffered by dogs that cough.”
“If you were already considering coughs, wouldn’t it have been better to include a function for tracking cardiomegaly as well?”
“I left it out because heart size varies depending on whether the dog is small, medium, or large. If you think it would be better to include it, I’ll add it.”
“I’m saying you could track it. Comparison with previous data should be possible, shouldn’t it?”
At the professor’s sharp tone, all the club members stiffened.
I seemed to be the only one who took that sharpness as something positive.
“Understood. We’ll add other lesions as well based on your later feedback, Professor.”
Giving specific feedback instead of showing indifference was a good sign.
He was not treating us like mere students; he was looking at us as people who were really trying to build a company.
“Try expanding the diagnostic scope into the musculoskeletal system later too. Things like patellar luxation.”
“Yes. Thank you for the feedback.”
“And don’t inflate the figures. It could actually hurt the product’s credibility.”
Then, out of nowhere, a different kind of feedback came flying in.
Don’t inflate the numbers?
“Which numbers do you think are inflated, Professor?”
At my question, Professor Shin smiled before answering.
“AUROC above 0.9. Accuracy at 94 percent. Those are numbers written with too much ambition.”
At that, I smiled as well and replied,
“Professor, those figures are from internal validation, but they are the performance our AI actually showed.”
At my words, the graduate students, who had been blank-faced until then, all stared at me with widened eyes, while the professor’s expression darkened slightly.
“Kim Siwon, I don’t know if you’re aware of this, but our lab is also developing lesion-diagnosis support software for dogs.”
It was skepticism born from expertise in the field.
He couldn’t readily believe that college students could demonstrate diagnostic performance that no company had been able to achieve.
“I am interested because this club has shown a level of preparation that doesn’t feel like a typical undergraduate group. That’s why I gathered everyone in the lab so we could discuss how to help each other. But acting this proud makes collaboration difficult.”
The professor looked more than a little offended, almost angry.
But letting ourselves be intimidated by that and saying, “We’ll verify the performance again,” would be foolish.
We weren’t complete outsiders, and we certainly didn’t need to bend ourselves into knots for an expert who was openly willing to help us.
“Professor, if you don’t believe our AI’s performance, you’re welcome to test it directly with your own data.”
On the contrary, it would be better to get him to actually use our AI through this opportunity.
“Test it?”
A hint of anticipation and curiosity finally appeared in the professor’s previously dark expression.
“Yes. It isn’t a formally approved product yet, but as far as I know, there shouldn’t be any legal issue with checking the data your lab already has or using it for academic purposes.”
“So you’re saying I should use it directly and see for myself? How accurate this AI really is?”
“That’s right.”
At my answer, the professor considered it briefly, then nodded.
The graduate students started writing something down at a rapid pace.
“I’ll review the legal issues and contact you.”
“Understood.”
“Doesn’t anyone in our lab have questions? Ask away.”
“Ah. I do have one question.”
The graduate students’ questions that followed were all quite basic.
What data had been used to train the software, how we planned to enter the actual target market, and so on.
Except for questions that couldn’t be answered because of product security concerns, I answered everything well.
“That part was planned by my colleague Ha Juwon, so would it be all right if she answered?”
To avoid looking like I was the only one doing all the speaking, I gave the other club members chances to answer here and there.
It was a way of proving that this was not the work of one competent person dragging a useless team along.
“Looks like everyone really worked hard, with no one freeloading.”
Thanks to that, both the professor and the graduate students started looking at our club with much softer eyes.
Professional enough to discuss business matters, but also genuinely passionate like college students, and above all, a group that worked together unusually well.
That was the image we needed to show them.
And it looked like the plan had worked.
“Wow, time really flew by. You all worked hard. I’ll be in touch again by email later.”
“Thank you.”
“Thank you.”
“We’ll see you again.”
The Q&A, which had stretched on for more than thirty minutes, wrapped up before it even reached an hour.
Thanks to the much more relaxed atmosphere than when we had started, we were able to smile as we opened the door and stepped back out.
“Wow. Professor Shin Dongjin is way scarier than I expected.”
“Didn’t he treat us less like college students and more like actual business partners?”
“Then that’s a good thing, right? He wasn’t laying into us just because he was annoyed, was he?”
“No way. Then why would he have gone out of his way to gather everyone in the lab?”
We, who had been quiet inside, started sharing our impressions the moment we left the school building.
Everyone seemed to have felt the pressure quite strongly.
“What if the professor intentionally runs all the hardest data and just tanks our accuracy?”
“It’s fine. He said his lab was already working on a diagnostic AI, too. We just have to be more accurate than that.”
“Wouldn’t he take offense?”
“Pride? Come on. He’s not even a specialist in this exact area, and it’s not like he’s trying to start a company.”
“Right. He might even decide to invest in our product.”
Even so, the feeling that things might really go well wasn’t mine alone.
The professor seemed genuinely interested in our product.
Since he had already done related research, it probably looked even more intriguing to him.
And with the notoriously exacting Professor Hwang Junhak as our advisor, he must have been viewing us through an even higher standard.
“I’ll show you how to install the product back at the club room. It’d be good for everyone to see it at least once.”
Then we’d just have to show him something better than that high standard.
---