Ammar Chaudhry

Aspiring Machine Learning Engineer with a solid foundation in Computer Science and a focus on using data-driven solutions to solve real-world challenges. Currently specialising in Data Science at the University of Leicester, I have developed strong technical skills in Python, machine learning, deep learning, and statistics through both academic work and hands-on projects.

During my scholarship in South Korea, I led a research project on air quality and public health, building a predictive model to explore the relationship between pollution and allergy symptoms. This international experience enhanced my skills in data analysis, predictive modelling, and cross-cultural collaboration.

Most recently, I have developed a deep learning-based train delay prediction system using RNN-LSTM models, integrating it into a live dashboard to support real-time forecasting—showcasing my ability to apply advanced ML techniques to time-series problems.

Combined with professional experience in hospitality, retail, and pharmaceuticals, I bring strong teamwork, time management, and adaptability—key traits for thriving in high-impact, fast-paced tech environments. I’m eager to contribute to intelligent systems that drive innovation and global progress.

Bachelor of Science: Computer Science (Final Year), Sept 2022 - Current

Turing Scheme Recipient | Sungkyunkwan University, South Korea | Jun 2023 – Jul 2023
Sponsored by the University of Leicester

• Academic Excellence: Achieved 92% in Data Analytics with Python and 90% in Cyber Security, excelling in a rigorous, data science-focused curriculum.
• Cultural Intelligence & Collaboration: Enhanced global perspective by working in a multicultural environment, honing teamwork and cross-cultural communication skills essential for diverse projects.
• Industry Engagement: Connected with leading scholars and data science professionals, expanding networks and gaining insights into the field.

Seoul Air Quality Analysis Project | Project Lead | Jun 2023 – Jul 2023

• Project Overview: Led a study to explore the relationship between air pollution (ozone, fine dust) and allergy symptoms across major cities in South Korea, aiming to identify patterns and forecast health impacts using real-world data.
• Data Analysis & Visualisation: Utilised Python libraries such as Pandas, NumPy, Matplotlib, and Seaborn for data cleaning, statistical analysis, and data visualisation, providing actionable insights into air quality trends.
• Predictive Modelling: Developed a predictive model with Scikit-learn to forecast air quality fluctuations and their potential health impacts, applying advanced machine learning techniques.
• Cross-Cultural Data Integration: Extracted and synthesised data from Korean government sources and open datasets, applying both technical and linguistic skills to manage data from diverse sources.
• Leadership & Collaboration: Led a multicultural team in executing the study, fostering collaboration across diverse perspectives, and ensuring timely project delivery.
• Presentation & Communication: Delivered findings to peers and professors, effectively communicating complex data insights and recommendations for future research.

Project Duration: 8 months

Overview:
Designed and implemented a real-time train delay forecasting system using deep learning. This end-to-end project applied a sequence-to-sequence LSTM model to historical rail and weather data, delivering live per-stop delay predictions through a deployed dashboard. The system was built with real-world impact in mind, combining advanced modelling, data automation, and user-facing deployment.

Key Features & Responsibilities:

Data Acquisition & Preprocessing:
Collected and integrated over 8,000 UK train journeys and 79,000+ meteorological records. Built a robust ETL pipeline using web scraping (BeautifulSoup) and the Google Sheets API, applying proxy rotation to manage API rate limits.

Deep Learning Model Development:
Built a Seq2Seq LSTM model using PyTorch to predict multi-step train delays. Designed the architecture to capture sequential and temporal dependencies between stops, dwell times, and exogenous variables (e.g. weather).

Model Optimisation:
Implemented hyperparameter tuning using Optuna with GPU-accelerated training, resulting in a test MAE of 1.16 min (arrival) and 1.22 min (departure). It outperformed traditional tree-based models in temporal generalisation while meeting sub-minute inference constraints.

Feature Engineering:
Engineered temporal, station-level, and operator-based features. Applied cyclical encoding, rolling delay statistics, and lag features to enhance predictive strength.

API & Dashboard Integration:
Deployed the model via FastAPI and integrated it into a live dashboard powered by real-time feeds. Delivered dynamic delay forecasts with sub-second latency, enabling proactive decision-making for passengers and operators.

Technologies & Skills Used:

• Deep Learning & Optimisation: PyTorch, Optuna, Seq2Seq LSTM, Scikit-learn
• Data Engineering: BeautifulSoup, Google Sheets API, Pandas, NumPy
• Deployment & Visualisation: FastAPI, Matplotlib, Seaborn
• Performance: GPU-accelerated training, hyperparameter tuning
• Project Management: GitLab, Agile

Outcome:
Delivered a scalable ML system for accurate, real-time train delay prediction. Conducted a comparative evaluation of four models (Random Forest and LSTM, with and without weather features), demonstrating that weather data increased MAE by +20% (LSTM) and +200% (RF)—highlighting delay propagation as the dominant factor.
The system is designed for scalability, with plans to extend coverage across all East Midlands Parkway services.
This project illustrates transferable skills in time-series modelling, full-stack ML deployment, comparative model evaluation, and AI product development for real-world impact.

Course Coordinator | Oxford Summer Courses (University Of Oxford) | Oxford, Oxfordshire | Jun 2024 - Jul 2024

• Managed a academic program for 16-17-year-old students, facilitating an authentic Oxford University experience.
• Conducted risk assessments and maintained safety protocols, ensuring participant well-being and compliance with institutional standards.
• Monitored student attendance and maintained detailed records to support program evaluation and improvement.

Key Transferable Skills: Project coordination, attention to detail, risk management, and stakeholder communication

Administrative Assistant | Travelodge Head Office | Thame, Oxfordshire | Apr 2022 - May 2022

• Exceeded daily targets in resolving customer complaints and inquiries in a fast-paced environment.
• Digitised a sentiment survey from the annual management meeting, structuring feedback for actionable analysis.
• Designed the 2022/23 rota for regional managers using advanced Excel, ensuring efficient scheduling.
• Collaborated with regional managers to streamline operations across UK Travelodge locations.

Key Transferable Skills: Data handling, analytical thinking, process improvement

Trainee Dispenser/Customer Advisor | Boots Pharmacy | Thame, Oxfordshire | May 2021 - Sep 2021

• Advised customers on medication usage, translating complex information into clear guidance.
• Prepared prescriptions with high attention to detail and maintained accurate medication records, supporting regulatory compliance.
• Collaborated with the team to manage stock levels and ensure operational efficiency.

Key Transferable Skills: Attention to detail, data management, communication, teamwork

Clitheroe Royal Grammar Sixth Form | Clitheroe, Lancashire

A-Level: Computer Science, Sept 2017 - Jun 2019

• Programmed a Virtual Trading software for a 100+ Business Studies Department students (Personal A-level Coursework Project)
• Achieved 93% for documentation and code corresponding to the A-level Coursework

A-Level: Mathematics, Sep 2018 - June 2020

• Developed quantitative skills and analytical thinking essential for data science applications.

A-Level: History, Sep 2018 - Jun 2020

• Conducted a critical analysis essay on the "Causes of World War One," enhancing research and analytical writing skills.