Annie Juliet Arul

Versatile software developer with expertise in Python, C++, Java, and JavaScript, complemented by proficiency in web technologies such as HTML and CSS. Skilled in object-oriented programming, data structures, algorithms, and advanced data processing using PySpark, Spark SQL, and Pandas UDFs. Demonstrated experience in machine learning with PySpark MLlib and deep learning using Inception V3 (MATLAB), with a strong focus on feature engineering and model evaluation. Adept at optimising file formats like Parquet and ORC for performance efficiency. Experienced in IoT development using NodeMCU (ESP8266) and Arduino, as well as integrating solutions with the Telegram Bot API. Committed to leveraging technical expertise to deliver innovative solutions while continuously advancing knowledge in cutting-edge technologies.

• Artificial Intelligence and Intelligent systems
• Scalable Big Data Analytics & Performance Optimisation
• Distributed Computing using Apache Spark
• Time-Series and Trend Analysis
• Risk Scoring and Analytical Metric Design
• End-to-End System Design and Development
• Problem Solving and Critical Thinking
• Communication and Team collaboration

• MATLAB Onramp - MathWorks
• Machine Learning Onramp - MathWorks
• Deep Learning Onramp - MathWorks
• Image Processing Onramp - MathWorks

Laser Security Protection System

IoT Project, Hindustan College of Arts & Science

Technologies: NodeMCU (ESP8266), Arduino IDE, Python, Telegram Bot API, L.DR Sensor, Laser Module

• Designed and developed an IoT-based laser security system with real-time intrusion detection and

Telegram alert notifications.

• Implemented sensor-based breach detection using laser and LDR with buzzer alerts and instant messaging.
• Enabled remote monitoring via Wi-Fi communication and ensured system reliability through end-to-end testing.

Multi-Label Genre Prediction Using Random Forest

Big Data & Machine Learning Project, University of East London

Technologies: PySpark, MLlib, Random Forest, Python, Pandas, Google Colab

• Built a scalable multi-label classification pipeline using Random Forest on large datasets in PySpark.
• Performed feature engineering, multi-label encoding, and class imbalance handling.
• Evaluated performance using Hamming Loss, Precision, Recall, and Macro F1-Score.

Weather & Climate Recognition Using Inception-V3, 

AI & Machine Vision Project, University of East London

Technologies: MATLAB, Deep Learning Toolbox, Inception-V3, Image Processing

• Developed a deep learning image classification model to recognize 11 weather conditions with 91.3% accuracy.
• Applied transfer learning, data augmentation, and real-time webcam-based prediction.
• Analysed performance using confusion matrix, precision, recall, and F1-score; authored a research paper on the project.

Design and Development of a Multi-Agent Machine Learning-Based Crop Recommendation System Machine Learning Project, University of East London

Technologies: Python, Scikit-learn, Pandas, NumPy

• Designed a multi-agent architecture for intelligent crop recommendation based on soil and environmental factors.
• Implemented data preprocessing, feature selection, and machine learning models for decision support.
• Improved recommendation accuracy through agent-based task separation and model evaluation.

Scalable Analytics on US Accidents Dataset Using PySpark

Big Data Analytics Project, University of East London

Technologies: PySpark, Pandas UDFs, Parquet, ORC, Spark SQL

• Performed large-scale analytics on the US Accidents dataset focusing on scalability and performance optimization.
• Applied window functions and time-series analysis to identify short-term accident severity patterns.
• Implemented Pandas UDFs and benchmarked performance against standard Python UDFs for risk score calculation.
• Compared Parquet and ORC file formats to evaluate storage efficiency and query performance.
• Optimized joins using broadcast join strategies to reduce data shuffle and improve execution speed.

Available to work up to 20 hours per week during term time (Flexible)