Skip to content

Open to UK placement opportunities

Mohith Thuraiyur Kirubanandham

MSc Data Science Student

Aspiring Data Scientist, AI & Machine Learning Enthusiast

University of SurreyUnited Kingdom

Turning data into insights. Building intelligence into solutions.

I'm an MSc Data Science student at the University of Surrey with an academic foundation in Artificial Intelligence and Data Science. I'm interested in turning complex data into meaningful insights and building intelligent solutions with machine learning and emerging AI technologies.

View projectsDownload CVContact me

Draw a digit

Draw any digit, 0 to 9Click and drag with your mouse.

AI sees

Loading the AI…

About me

From an undergraduate degree in AI to postgraduate data science in the UK.

I studied Artificial Intelligence and Data Science at Rathinam College of Arts and Science in Coimbatore, where my projects ranged from fake news detection and stroke-risk prediction to an encrypted vehicle black box. In 2026 I moved to the UK to begin an MSc in Data Science at the University of Surrey.

I'm most interested in analytical problem-solving: understanding a complex dataset, finding the patterns that matter, building predictive models and checking whether the results hold up statistically. I want to apply that to real problems, especially in healthcare and safety.

Alongside my studies I build independent projects, from a market drawdown alert system to backend services, to practise the engineering side of working with data. I'm looking for a placement in data science, data analytics, machine learning or AI in the UK.

Technical skills

Tools and topics I use or am learning, grouped by area. These aren't proficiency ratings: where an area is used in a project, the project is linked.

Programming

Python, SQL, R, C++, JavaScript

Used in NIFTY 50 Drawdown Alert System, Asynchronous Analytics Job API, Bookshop Microservices API and RK Creators & Builders Website

Data analysis

pandas, NumPy, Data cleaning, Exploratory data analysis, Statistical analysis

Used in Stroke Prediction System

Data visualisation

Power BI, Matplotlib, Seaborn, Excel

Learning through my MSc and self-study

Machine learning and AI

Supervised learning, Unsupervised learning, Predictive modelling, scikit-learn, CNNs, Natural language processing, Large language models, Retrieval-augmented generation

Used in Fake News Detection Model and Stroke Prediction System

Databases

MySQL, SQLite, Relational database concepts, Database management

Used in Bookshop Microservices API and NIFTY 50 Drawdown Alert System

APIs and backend

Flask, REST API development, Microservices, Node.js, Serverless functions

Used in Asynchronous Analytics Job API, Bookshop Microservices API and RK Creators & Builders Website

Cloud and infrastructure

Cloud computing, Internet-scale systems, Docker, CI/CD with GitHub Actions

Used in NIFTY 50 Drawdown Alert System and RK Creators & Builders Website

Development tools

Git, GitHub, VS Code, Jupyter Notebook, Google Colab, Linux / Ubuntu

Learning through my MSc and self-study

IoT and embedded

NodeMCU (ESP8266), Arduino, Sensor integration

Used in Secure Vehicle Black Box

Featured projects

Coursework, independent and undergraduate projects, strongest first. Open a project to read the full case study.

Showing 8 projects

Independent project2026

NIFTY 50 Drawdown Alert System

A monitoring service that tracks the NIFTY 50 index and sends an alert when it falls set percentages below its rolling 52-week high.

  • Python
  • yfinance
  • SQLite
  • Docker
  • Telegram Bot API
  • pytest

View code for NIFTY 50 Drawdown Alert System on GitHub (opens in a new tab)

Problem
Market drawdowns build up over days or weeks, and checking an index by hand makes it easy to miss the moment a decline crosses a level that matters.
Objective
Watch the NIFTY 50 during market hours and send one alert per drawdown threshold by Telegram or email, without ever touching a trading account.
Read case study
My contribution
Designed and built the data pipeline, alert logic, notifications and deployment setup.
Approach
  • Prices come from yfinance through a pluggable data provider and are validated before use.
  • The service computes the drawdown from the rolling 52-week high and checks it against thresholds of 10% to 50% in 5-point steps (configurable).
  • Alerts go out through Telegram and SMTP email, with SMS and push left as stubs for a provider to plug in.
  • It runs as a systemd service, a cron job, or two Docker Compose containers: the monitor and a Telegram assistant that answers status questions.
Key technical decisions
  • Each threshold fires once and only re-arms after the drawdown recovers within a 2% buffer, so small swings don't repeat alerts.
  • Alert state is stored in SQLite, so a restart never resends an alert.
  • Polling is gated by market hours, weekends and an exchange-holiday file, and stale prices are rejected.
  • Monitoring only: the system has no broker or demat access by design.

MSc coursework2026

Asynchronous Analytics Job API

A Flask API that accepts long-running analytics jobs, runs them in the background and lets clients poll for status and results.

  • Python
  • Flask
  • REST

Coursework version: processing is simulated and jobs are held in memory.

View code for Asynchronous Analytics Job API on GitHub (opens in a new tab)

Problem
Analytics jobs such as MapReduce or Spark runs take too long for a single request and response, so clients need a way to submit work and check back later.
Objective
Implement the asynchronous job pattern for an analytics service: submit a job, track its status, then fetch results when ready.
Read case study
My contribution
Implemented the service as part of University of Surrey coursework (COM3033).
Approach
  • POST /jobs validates the request and returns 202 Accepted with a job ID.
  • A background thread moves each job from QUEUED to RUNNING to SUCCEEDED.
  • GET /jobs/<id> returns the status, and GET /jobs/<id>/results returns 409 until the job has finished.
  • /health and /metadata describe the service and the engines it supports.
Key technical decisions
  • Specific JSON error codes (MISSING_FIELDS, INVALID_ENGINE, JOB_NOT_FOUND, RESULT_NOT_READY) instead of generic failures.
  • The engine field is limited to mapreduce or spark, matching the platforms the service is designed around.

BSc academic project2026

Secure Vehicle Black Box

An IoT black box prototype that records vehicle sensor data, encrypts every packet and stores it for post-incident analysis.

  • NodeMCU (ESP8266)
  • Arduino C/C++
  • Python
  • RC6 encryption
  • ThingSpeak

Prototype. Tamper detection, event replay and automated reports are described as design goals but are not yet in the published code.

View code for Secure Vehicle Black Box on GitHub (opens in a new tab)

Problem
After a road incident, investigators and insurers need sensor records they can trust have not been altered.
Objective
Capture motion, location and environmental readings from a vehicle and keep them in an encrypted log for later analysis.
Read case study
My contribution
Developed the NodeMCU firmware and the Python encryption and decryption scripts in this repository.
Approach
  • A NodeMCU reads an MPU6050 accelerometer and gyroscope, a NEO-6M GPS module and a DHT11 temperature and humidity sensor, with a 16x2 LCD for local display.
  • A Python script polls the device every 5 seconds, encrypts each reading with RC6 using a 128-bit key and appends it to a CSV log.
  • A second script decrypts the log for analysis, and ciphertext can also be sent to ThingSpeak dashboards.
Key technical decisions
  • Symmetric RC6 encryption with a shared 128-bit key between the logging and decryption scripts.
  • A push button and a potentiometer simulate braking and speed, so the prototype can be tested on a bench.

MSc coursework2026

Bookshop Microservices API

Three Flask services, a gateway, a catalogue and an orders service, that work together over HTTP to run a small online bookshop.

  • Python
  • Flask
  • SQLite
  • REST
  • Requests

View code for Bookshop Microservices API on GitHub (opens in a new tab)

Read case study
Problem
Splitting an application into services creates new failure modes: a service can be slow or down, and stock has to stay consistent across services.
Objective
Build a gateway-fronted system where each service owns its own data and failures are reported clearly.
My contribution
Implemented the gateway, catalogue and orders services as University of Surrey coursework.
Approach
  • The gateway exposes /api/books and /api/orders and forwards each request to the service that owns it.
  • The catalogue service keeps books and stock in SQLite, and the orders service keeps orders in its own SQLite database.
  • Placing an order checks the book with the catalogue, reduces its stock, then saves the order.
Key technical decisions
  • Every call between services has a 2 to 3 second timeout, so one slow service can't hang the system.
  • Clear status codes: 502 when the gateway can't reach a service, 503 when the catalogue is down, 409 on insufficient stock.
  • The gateway's /health endpoint combines the health of every service and reports degraded with a 503.
  • Ports, service URLs and database paths come from environment variables, ready for containers.

BSc academic project2023–2026

Fake News Detection Model

A convolutional neural network that classifies news articles as real or fake from their text.

  • Python
  • CNN
  • Deep learning
  • NLP

Code not yet published.

Read case study
Problem
Misinformation spreads faster than people can check it by hand, so automated screening of news text is a useful first filter.
Objective
Train a deep learning model to separate real and fake news using the article text.
My contribution
Developed the model as a BSc academic project.
Approach
  • Preprocessed the article text for the model.
  • Built a CNN architecture for text classification.
Key technical decisions
  • A convolutional network, which picks up local word patterns in text, as the classifier.

BSc academic project2023–2026

Stroke Prediction System

A machine learning model that estimates a patient's stroke risk from their attributes.

  • Python
  • scikit-learn

Academic project, not for clinical use. Code not yet published.

Read case study
Problem
Spotting people at higher risk of stroke early can support earlier medical follow-up.
Objective
Predict the likelihood of stroke from patient attributes using supervised learning.
My contribution
Developed the analysis and model as a BSc academic project.
Approach
  • Explored the patient attributes with exploratory data analysis.
  • Selected features for the model.
  • Built the predictive model with scikit-learn.
Key technical decisions
  • Framed as early risk identification only, not diagnosis.

Independent project2026

RK Creators & Builders Website

A website and admin panel for a family-run construction business in Thuraiyur, Tamil Nadu.

  • HTML
  • CSS
  • JavaScript
  • Node.js
  • Netlify Functions
  • Netlify Blobs
  • GitHub Actions

View code for RK Creators & Builders Website on GitHub (opens in a new tab)

Read case study
Problem
The business needed an online presence where customers can see its services and past work and send enquiries.
Objective
Ship a fast static site with a working enquiry form and a private panel for managing projects and enquiries.
My contribution
Built the site, the serverless functions and the deployment pipeline.
Approach
  • Static HTML, CSS and JavaScript pages for home, about, services, process, projects and contact.
  • Netlify Functions handle enquiries, admin login, and project and enquiry management, with data stored in Netlify Blobs.
  • A GitHub Actions workflow deploys to Netlify on every push to main.
Key technical decisions
  • The enquiry endpoint uses a honeypot field, per-IP rate limiting and input validation to cut spam.
  • New enquiries are emailed through Nodemailer.
Challenges
  • The CI deploy failed on a platform-specific lockfile conflict; fixed by installing only the functions' dependencies in the workflow.

Team project2025

Medical Inventory Management

A Salesforce app for tracking medical suppliers, products, purchase orders and expiry dates.

  • Salesforce
  • Apex
  • Lightning
  • Flows

The repository holds the team's project report.

View code for Medical Inventory Management on GitHub (opens in a new tab)

Read case study
Problem
Medical stock has to be tracked against suppliers and expiry dates so that expired items are not used.
Objective
Build a Salesforce Lightning app that manages suppliers, products, purchase orders and inventory transactions.
My contribution
Team member in a Naan Mudhalvan programme project; the work was shared across the team.
Approach
  • Custom objects for products, purchase orders, order items, inventory transactions and suppliers.
  • Validation rules, page and compact layouts, profiles, roles and permission sets.
  • A flow that updates actual delivery dates and an Apex trigger that totals order-item amounts per purchase order.
Key technical decisions
  • Purchase order totals are calculated by an Apex trigger so they stay correct when order items change.

Research interests & areas of exploration

Areas I'm reading about and want to work in. These are interests, not completed research.

  • Artificial intelligence and machine learning

    Exploring intelligent systems, predictive modelling, machine learning algorithms and practical AI applications.

  • Healthcare and AI

    Interested in AI-powered healthcare applications, diagnostic assistance and data-driven healthcare research.

  • Vehicular safety and intelligent systems

    Interested in applying AI, sensor data and computational methods to improve vehicle safety and accident analysis.

  • Natural language processing

    Exploring NLP, language models, text processing and intelligent information retrieval.

  • Retrieval-augmented generation

    Interested in building knowledge-grounded AI systems that combine document retrieval with language models.

  • Data science research

    Exploring statistical analysis, experimental design, reproducibility, data-driven discovery and research methods.

Academic journey

How my MSc builds on an undergraduate foundation in artificial intelligence and data science.

MSc Data Science

University of Surrey, Guildford, United Kingdom

2026 – 2027In progress

MSc learning areas

Modules I'm studying this year, listed as learning areas rather than qualifications.

  • Statistical Data Science
  • Database Systems and Business Intelligence
  • Cloud Computing
  • Computational Intelligence
  • Engineering Internet-Scale Systems
  • Business Analytics with Data Visualisation
  • Machine Learning for Data Science
  • Artificial Intelligence and AI Programming
  • Natural Language Processing
  • Research Ethics and Security

Coursework write-ups

Write-ups of MSc assignments will appear here as modules finish, each with the problem, dataset, method, results and an evaluation of what worked.

Lab projects from this term are already in Projects.

BSc Artificial Intelligence and Data Science

Rathinam College of Arts and Science, Coimbatore, India

2023 – 2026Completed

Covered: Python, deep learning, data analysis, database management systems, data visualisation, IoT, predictive analytics.

English language: IELTS 6.5

Currently building & exploring

What I'm working on alongside my lectures. Last updated October 2026.

  • Strengthening Python and SQL
  • Improving statistical and analytical skills
  • Developing practical machine learning projects
  • Building data visualisation dashboards
  • Exploring cloud-based data systems
  • Experimenting with LLMs and RAG architectures
  • Improving software engineering and GitHub practices
  • Preparing for technical placement interviews

Certifications & professional development

Only certificates I've completed and can verify are listed here.

No certifications listed yet

Completed certificates will appear here with the issuer, completion date and a verification link.

Courses in progress

None outside my degree right now. My current study is the MSc Data Science programme, listed under Academic journey.

Academic & independent experience

Coursework, academic and independent work. None of it is paid employment.

  1. 2026 – present

    University of SurreyMSc coursework

    Building backend and data systems in coursework, including an asynchronous analytics job API and a bookshop microservices system.

  2. 2026

    Self-directed projectsIndependent

    Built a NIFTY 50 drawdown alert system with Docker and tests, and a website with serverless functions for a family-run construction business.

  3. 2023 – 2026

    Rathinam College of Arts and ScienceBSc academic

    Academic projects in fake news detection with a CNN, stroke-risk prediction and an encrypted IoT vehicle black box.

  4. 2025

    Naan Mudhalvan programmeTeam project

    Contributed to a team-built Salesforce app for medical inventory management.

Beyond technology

  • Fitness and strength training
  • Trekking and outdoor exploration
  • Technology and innovation
  • Continuous learning

Let's connect and build something meaningful.

I'm looking for data science, data analytics, machine learning and AI placements in the UK. Email is the quickest way to reach me.

This opens your email app with the message filled in.