Cyber Security undergraduate · Dubai

Curious mind.
Secure future.

Hey, I’m Tharish J. I’m exploring the intersection of cyber security, code and AI — turning what I learn into things I can build.

Explore my work ↘LinkedIn ↗

An explorer’s mindset

< learn. build. repeat. />

CYBER SECURITY / AI01 — PERSONAL PORTFOLIO

⌘

University of Wollongong in DubaiUndergraduate student

◎

Cyber SecurityAcademic specialization

↗

Artificial IntelligenceSamsung Innovation Campus · 2025

01 / Featured team project

From market data
to machine learning.

Samsung Innovation Campus
Artificial Intelligence · Group 33 · 2025

REPORTED CONFUSION MATRIXTEST SET

1,389DOWN → DOWN

1,097DOWN → UP

898UP → DOWN

966UP → UP

Stock data→24 inputs→UP / DOWN

Source: Team 33 project presentation

Six-person academic project

Stock Price Movement Prediction Using Random Forest

Can historical stock data help predict tomorrow’s direction? Our team built a Python pipeline that turns price and volume data into features for an UP / DOWN classifier.

Pythonpandasscikit-learnGoogle Colab

54%Reported test accuracy

0.549Reported ROC-AUC

24Model inputs

250Decision trees

Results from the submitted report and presentation, using an 80/20 stratified random split. This evaluation does not establish performance on future market data.

My contribution

Framing the problem and solution.

Introduction, problem definition and proposed solution, as credited in the final project report.

Built together / Group 33

Tharish J · Naman Chawla · Dina Abdelnaser Mohammed · Raia Mohammed Gyeye · Nicelin Printo · Mariam Lulu Mohammed

01 How the project works

DATA / 01

Prepare the market history

Sort daily stock records by company and date, then clean missing and non-finite values.

FEATURES / 02

Give the model context

Combine price and volume with returns, volatility, moving averages, RSI, MACD and Bollinger Bands.

MODEL / 03

Classify and evaluate

Train a Random Forest and examine accuracy, class performance, a confusion matrix and ROC-AUC.

02 Results and what we learned

The reported accuracy was modest. The ROC-AUC of 0.549 also sits close to 0.5, highlighting how difficult next-day direction is to predict from these inputs. The project is an academic research prototype for learning and analysis.

A useful next step is chronological or walk-forward validation, alongside richer data and alternative models.

Original Team 33 results slide showing 54 percent test accuracy, 0.549 ROC-AUC, confusion matrix and ROC curve.

01 / PDF ↗Read the reportFull methodology and analysis 02 / PDF ↗View the presentationThe project in 10 slides 03 / PYTHON ↓Download the codeThe submitted model pipeline 04 / ZIP ↓Download the datasetOriginal CSV, compressed

02 / My toolkit

Explore my learning.

Choose an area to see the tools and interests behind my work.

From data to an experiment.

My team project brought together data preparation, feature engineering and model evaluation.

PythonpandasNumPyscikit-learnMatplotlibSeaborn

03 / Beyond the screen

Always a student.
Always curious.

I’m Tharish Jayachandra Babu, from Tamil Nadu, India, and studying in Dubai. I enjoy programming, practical learning and seeing how technology works beneath the surface.

Outside my studies, you’ll often find me gaming. I like a good challenge — whether it’s in a game or a piece of code.

Get to know me on LinkedIn ↗

CURRENT STUDIES

University of Wollongong in Dubai

BSc student specializing in Cyber Security.

2025 / CERTIFICATE

Samsung Innovation Campus

Artificial Intelligence Course · Samsung Gulf Electronics

2025 / TEAM PROJECT

Applying AI to stock data

Exploring classification through a practical research prototype.

04 / Start a conversation

Good ideas start
with a hello.

Have a project idea, a learning opportunity, or a shared interest in technology? Let’s connect.

Email me ↗

nbtharish@gmail.com

Connect on LinkedIn ↗

Curious mind.
Secure future.

From market data
to machine learning.

Stock Price Movement Prediction Using Random Forest

Framing the problem and solution.

Prepare the market history

Give the model context

Classify and evaluate

Explore my learning.

From data to an experiment.

Always a student.
Always curious.

University of Wollongong in Dubai

Samsung Innovation Campus

Applying AI to stock data

Good ideas start
with a hello.

Stock direction, explored through AI.

My contribution

The workflow

Curious mind.Secure future.

From market datato machine learning.

Stock Price Movement Prediction Using Random Forest

Framing the problem and solution.

Prepare the market history

Give the model context

Classify and evaluate

Explore my learning.

From data to an experiment.

Always a student.Always curious.

University of Wollongong in Dubai

Samsung Innovation Campus

Applying AI to stock data

Good ideas startwith a hello.

Curious mind.
Secure future.

From market data
to machine learning.

Always a student.
Always curious.

Good ideas start
with a hello.