About Me

As a person

I am a critical thinker with strong analytical skills and a passion for Information Technology and computer science—fields as vast and deep as the ocean. I love expanding my knowledge in IT, which is also my hobby. In my free time, I enjoy nature or immerse myself in reading to broaden my understanding of science and technology.

Things I Can Do

Dedicated and fast learner with a strong background in systems and network engineering. Excellent in coding, with hands-on experience in both private and public cloud computing. Currently expanding expertise in deep learning.

AI and Machine Learning (ML)

While pursuing my Data Science studies at Halmstad University in Sweden, I undertook the following projects.

NegoIT - Inward Revenue Forecast

This project was conducted in collaboration with a valued client of NegoIT, a wholesale distribution company engaged in both product importation and manufacturing. Maximizing inward revenue is crucial for effective import planning and ensuring product quality to optimize profitability. However, accurately forecasting inward revenue presents challenges due to the nuanced nature of sales dynamics. These include varying credit lines for different clients, occasional adjustments based on promotional activities, and multifaceted influences such as seasonal patterns, historical purchasing data, and external factors like weather conditions.

To address these complexities, I developed two distinct solutions: Long Short-Term Memory (LSTM) and Autoregressive Integrated Moving Average (ARIMA). Following rigorous evaluation, the LSTM model, augmented with an autoencoder, emerged as the optimal choice. This final implementation consistently achieved over 92% accuracy in predicting inward revenue across diverse scenarios.

Halmstad University - Fuel Consumption Analysis and Anomaly Detection for a Vessel in Short-Sea Shipping

In a project centered on a cargo ship's data between Vaxholm and Sodra in Stockholm, I employed various techniques to identify anomalies in fuel consumption. Initial clustering with KMeans revealed two distinct clusters and outliers. Anomaly scores were computed using LOF, leading to a focused analysis of anomalous records. Further exploration involved Dynamic Time Warping (DTW) on the engine fuel rate timeseries. For root cause analysis, feature correlation was examined using xgboost, and SVM successfully predicted the ship's route based on weather factors, directly influencing fuel efficiency.

Halmstad University - Apache Spark Big Data Parallel Programming

The project focused on acquiring hands-on experience with big data and selecting the optimal cluster for the specified dataset. Multiple Apache Spark clusters were established on Google Cloud Platform (GCP) and trained using PySpark. The project involved extensive data preprocessing, including cleaning, addressing imbalances, filtering, and transformations through pipelines. Machine learning techniques, such as Logistic Regression with K-Fold Cross Validation, Decision Tree, Random Forest, SVM, and Naïve Bayes models, were trained to further enhance the project's analytical capabilities.

Halmstad University - cGAN

In a university Deep Learning project, I utilized PyTorch with CUDA to design and train a Conditional Generative Adversarial Network (cGAN). The goal was to predict the last frame of an image sequence given the initial n frame. Employing multiple layers of CNN, Trans-CNN, FC, and incorporating techniques like dropouts, skip connections, and data augmentation, I achieved exceptional model performance. Extensive experimentation with optimizers and activation functions led to outstanding results, earning the project the highest grade.

My IoT Journey: Innovations and Applications

Exploring IoT: A Hobby Turned Passion

During the 2018 to 2019 period, I embarked on an exciting journey into the world of Internet of Things (IoT) as a hobby. This exploration led me to develop numerous innovative applications, including SMS gateways, weather stations, biometric systems, robotic applications, and irrigation control systems. Leveraging my strong proficiency in PHP, JavaScript, and MySQL, I created web applications to manage and control most of these IoT projects.

Image 1 Image 2 Image 3 Image 4 Image 5 Image 6 Image 7 Image 8

Highlight Project: Soil Moisture Irrigation System

One of my most fascinating projects was the soil moisture irrigation system designed for home gardening. This system could connect up to four capacitive soil moisture sensors and automatically manage irrigation based on soil moisture levels, air temperature, and humidity.

Key features of this project include:
  • Automatic Sensor Fault Handling: The system can detect and handle soil moisture sensor failures, ensuring continuous operation.
  • Initial Setup: When powered on for the first time, the device creates an open access point and hosts a small internal web application. Users can connect via a mobile phone to configure the home Wi-Fi settings and device ID.
  • Web Application Integration: Once connected to the internet, the device can be accessed through a web application. Users can adjust settings, manually trigger irrigation (Water Now function), and generate reports.
  • Firmware Updates: The system can upload report updates to a web server and store them in a database. On the next API call, the device automatically downloads and installs new firmware.
  • PCB Design: I have a working knowledge of electronics and have designed many circuits and antennas. During this project, I designed the electronic circuits and PCB motherboard and used JLCPCB in China to print the board.

Innovative Application: SMS Gateway

Another notable project was the SMS gateway system. This application was designed to send messages to a web-based API, which are then picked up and dispatched by multiple SMS gateways. The system operates on a first-come, first-served basis and includes robust fault tolerance mechanisms. Key features include:
  • Queue Management: Messages are placed in a queue and dispatched in the order they arrive.
  • Redundancy and Fault Tolerance: If a device fails to send an SMS, another device in the network will dispatch the message after a dead letter queue timer expires, ensuring no message is lost.

Conclusion

My journey into IoT has been a rewarding experience, allowing me to combine my programming skills with innovative technology to create practical solutions. From managing home irrigation systems to ensuring reliable SMS communication, these projects showcase the potential of IoT to simplify and enhance everyday tasks.

Software Engineering

One of my primary projects involved developing a large-scale, PHP-based sales system. This comprehensive application includes modules for sales, inventory, availability search, item transfer, management, top management, purchase orders, marketing, decision making, reporting, settings, HR, and finance. Additionally, the system features an API to integrate with external SMS gateways and e-commerce sites. Key technologies and features include:
  • Front-end Optimization: Utilized Ajax extensively to enhance application speed and responsiveness.
  • Database Management: MySQL serves as the database, hosted on AWS RDS with cross-region replication for redundancy.
  • AWS Integration: Leveraged various AWS services for a robust and automated platform:
    • EC2 with Autoscaling: Ensures scalability and high availability.
    • Application Load Balancer (ALB): Distributes traffic efficiently across multiple instances.
    • Elastic File System (EFS) and S3: Provide scalable and reliable storage solutions.
  • Linux Hosting: The application is hosted on a Linux platform, ensuring a stable and secure environment.
This project exemplifies the integration of multiple technologies and services to create a fully automated and efficient sales system.

Network Engineering

I have 13 years of experience in the network and systems sector, during which I have successfully implemented numerous technology revamps and upgrades. I prioritize meticulous planning to ensure zero system downtime. As a highly trained Microsoft and Cisco engineer, I have undergone extensive classroom training to enhance my expertise.

Throughout my career, I have worked full-time in systems and network roles at three different companies. During this time, I have deployed and managed various networking devices. Some of the key devices I have worked with include:
  • Cisco 4500, 9300, 1841, WLC 2500, ASA 5500, AIR-CAP2702, AIR-CAP1702
  • Palo Alto PA-3020, PA-820
  • Fortinet FG-601E
  • Brocade switches ICX6610, ICX7250
  • IBM SAN switches
  • Storage: HP P2000, IBM DS4700, IBM V3700, Lenovo DE2000H
  • Many Microsoft applications including AD, Exchange
My comprehensive experience and dedication to maintaining optimal system performance have consistently driven successful outcomes in my projects.
Image 1