Aina plugs in the extension cable and switches on her laptop. The screen flickers to life, and with a few quick taps, her project is up and running — the interface glowing softly under the hall lights.
She takes a step back to admire the setup. Her banner reads:
"Cat Language Translator — Using AI to Understand Feline Emotions"
Posters showing waveform diagrams, emotion tags, and screenshots of the app frame the desk. A neat little card reads:
Developed by: Aina Raza
Department of Computer Science
Her laptop is running, the app interface neatly opened. A tablet beside it shows the waveform visualizations. A small speaker is ready to play the meows, and she's printed emotion categories in bright color-coded cards to place beside each output: Happy, Warning, Defensive, Mating, and so on.
Aina exhales slowly, hands clasped in front of her. She isn't just presenting a project. She is presenting her pain, her late nights, her loneliness — all packaged in code and courage.
Within minutes, visitors start entering the hall — students, teachers, alumni, and some industry professionals with badges clipped to their collars.
The first group stops at her table, a man with thick glasses and a clipboard.
"What's this?" he asks, pointing to the screen.
Aina stands straight, smiles confidently, and begins, "It's a Cat Language Translator App. It uses machine learning to analyze and classify a cat's meow into emotional categories like 'Hungry', 'Angry', 'Resting', 'Happy', and so on."
The man raises an eyebrow. "You're saying a machine can understand cat emotions?"
Aina nods. "Not perfectly — but with enough data and training, yes. I've trained it on ten categories using real vocalizations, and the model gives a predicted label based on the input meow."
Another visitor leans in, intrigued. "How do you collect the data? Cats aren't exactly the most cooperative subjects."
Aina chuckles softly. "I use publicly available datasets at first, then collect some with the help of vets and pet owners. It's challenging, but... worth it."
The man with the clipboard taps his pen on his lip. "Show us."
Aina clicks the demo button. A recorded cat sound plays, and within seconds, the app displays: Predicted Mood: Warning
A murmur of surprise spreads around her table.
"This is... impressive," says a woman with a scarf, examining the UI. "And it's mobile-friendly too?"
"Yes," Aina nods, her eyes shining now, "I developed it in Flutter. The backend is integrated with Firebase, and I use TensorFlow Lite for the ML model."
They all exchange looks. One whispers, "She did this alone?"
Aina hears it but pretends not to. Her chest swells just a little with pride. This moment—this rush—is everything she has worked for.
A younger student pipes up, "Wait, this can actually help pet owners, right? Like for medical alerts or behavior training?"
"Exactly," Aina says, clicking to the next slide. "It could help in early detection of distress or aggression, especially when owners don't notice subtle signs. There's potential for smart collar integration, too."
The woman visitor gives a slow clap. "This is real innovation."
Aina smiles wider. She doesn't feel lonely for the first time in a long while. She feels... seen.
Just then, someone clears their throat softly.
Aina looks up to see a man in his early thirties, in a grey blazer and glasses — his ID card labels him as Visir Khan – Industry Judge (AI Startups).
"Mind if I ask a few questions?" he says politely.
"Of course, sir," Aina replies.
He steps forward, eyes flicking over the setup. "You built this alone?"
"Yes, sir."
"Interesting," he murmurs, then points at the waveform chart. "What features did you extract for classification?"
"I use MFCC — Mel-frequency cepstral coefficients — to capture the timbral features of the cat sounds. The model is a CNN-based classifier, trained on ten mood categories."
"CNN," Visir repeats, impressed. "Did you augment your dataset?"
"Yes. The original dataset was limited, so I applied time-stretching, pitch shifting, and background noise injection. But only on training data — not on test sets."
His eyebrows raise. "Good. That's a common mistake among undergrads."
She smiles faintly. "I made it once… then re-trained everything."
He chuckles. "Even better. Shows you care about results."
Then he asks, "And deployment?"
"I use TensorFlow Lite. Integrated it into a Flutter-based mobile app. It runs offline too."
Visir folds his arms, lips quirking in amusement. "Aina Raza… you've done something unusual here."
"Thank you, sir."
He takes a slow breath, then says, "You know… there's a startup in Eskişehir trying to recognize emotion in animals. Dogs, mostly. But they're nowhere near product-level yet."
Aina's eyes widen. "Really?"
"Yes," he nods, then leans a little closer. "If you're serious about this, I'd suggest you don't stop here. There's a market for this."
"I want to continue," she says firmly, surprising even herself with the certainty in her voice.
Visir smiles. "Then I'll be watching your stall closely today. Good work, Miss Raza."
And with that, he moves on, leaving Aina stunned — and glowing.
People keep coming in waves. Some try the app, some laugh at the predictions, and others whisper with surprise. But Aina stands proud, confident, grounded.
She doesn't notice, however, that someone has been watching her through the crowd.
Not Visir.
Not a classmate.
A stranger. With unreadable eyes.
And he isn't looking at the app.
He is looking at her.