#DigitalDefence – A Tribute To The Teams

It was fantastic to see / hear / participate in the closing ceremony of the #DigitalDefence Hackathon 2020. If you want to check the whole ceremony including some of locknotes, check it out here.

#DigitalDefence Hackathon 2020 Closing Ceremony recorded by Hackmakers (https://hackmakers.com)

It was great to see who won but also from the judging perspective, who else was in the Top 11 (yes 11, not 10) where we worked with our executive team including Cherie Ryan, Vice President at Oracle and our Regional Managing Director of Australia and New Zealand to pick the winners.

We were honoured that we were able to help so many of the top teams to execute and demonstrates what we are capable of as when we work together with positivity. Much of I’ve learnt over the years in these communities, you get 10x whatever you put in. This is not to be selfish. The attitude is to ensure that we always push hard to (sustainably) give. “Sitting back and waiting for things to happen will not create the opportunities.” Hence, having the ability to actively contribute in a positive environment will build capability, confidence and the community (this is where the 10x comes in).

It definitely showed that our mission statement is built into our culture.

The rest of the article is about the teams and a tribute to them. These aren’t in any specific order, or chosen for a specific reason. Only except that we had the chance to collaborate in some way with these fantastic people and have enough here to represent their work and outcomes.

As a call to action and in your own way, here is an opportunity to build and be part of this ever growing community. Please give if you can. One of these teams may be able to help you to connect better with the problem, with the idea, with the people and organisations that supported them – endless possibilities.

Reach out and engage.

NOTE: These projects and presentations were delivered in 2.5 days. No one expects perfection in this timeframe. The purpose is to demonstrate capability and provide insights into the possibilities. That being said there are many learnings that came from this process.

NOTE: I’ve also tagged as many of the team members as I can so you can reach out. If there are others, please let me know via LinkedIn on my profile (HERE)

Team TARDIS – 5th Placed Finalist


The Team:
Arya Anghan, David Sarkies, Mohammed Ilyas Ahmed, Ritu Kumari., Szabolcs (Szasza) Palmer

The Problem / Challenge:
The team was focusing on the “Cyber security and exploitation mitigation” challenge with the specific area being the active, real-time exploitation prevention.

The Solution:
They built a python-based solution using the Oracle Cloud Infrastructure. When a file is uploaded, the solution registers the uploader’s IP address; checks with the IBM X-Force Exchange database via its API to determine if either the IP address or the file content hash is malicious and deletes the uploaded file if it is. This happens all in real-time and the verification happens within a couple of milliseconds from the upload.

Identity Crisis – 3rd Placed Finalist


The Team:
Suchet Singh, Deepit Amin, Astha Chauhan, Atuf Abdul khader, Shea Mithala Parambath, Khan Nomaan

The Problem / Challenge:
The team was focusing on the “Deep Fake” challenge with the specific area being the detection of fabricated videos.

The Solution:
They used the encoded pixel based area of which our face is being recognised in the eye and lip region. Making use of a machine learning training model, it is able to recognise and detect of the image is real or fake.

Black Cap – Winner


The Team:
Tarek Chaalan, Imad Mehmood, Andleeb Raja, Robert Dzudzar, Neda AfzaliSeresht, Khai Fahmi Zaki, Asif Rasool

The Problem:
The team was focusing on the “Anomaly Detection” challenge with the specific area being the money laundering scenario based upon the “Charlie Crash Repairs” story.

The Solution:
By analysing the data from the business which includes the account information and transaction information, they located and analysed multiple anomalies within the given business. Anomalies that were found included suspicious large transactions with over $90K, self-self transactions, circular transactions, unusual amount of money spent on meals and people having double accounts. Utilising different tools like python, Neo4J, Graphistry and Tableau hosted on Oracle Cloud Infrastructure, they showed this anomalous transaction visually by creating network graphs and interactive visualisations.

DineDetect – 4th Placed Finalist


The Team:
Ayaz Mujawar, Jainish Shah, Revathy Sivaraman, Hanifa Baporia

The Problem / Challenge:
The team was focusing on the “Anomaly Detection” challenge with the specific area to support restaurant owners who are facing fraud in terms of complains and refunds. Some customers cancel online food orders by giving fake complaints and ask for refunds after consuming the food. This is costing massive losses to a restaurant owner.

The Solution:
They use a clustering technique to detect the customer demographic information like location, account number and phone number and based on that, they mark the customers as either a fraudulent and non-fraudulent customer.

The Seekers – Runner-Up


The Team:
Ka Pui (Joey) Yuen, William Tin, Nauman Akram, Thulasiram Nagam, Ali Haider, Muhammad Kafeel, Maria Sheikh, Rabi Siddique

The Problem / Challenge:
The team was focusing on the “Anomaly Detection” challenge with the specific area being anomalies in network traffic.

The Solution:
The team focused on a range of features from incoming ports to destination ports, state, packets and services. From there, these features were used to predict given some feature values if traffic is anomalous or normal. During the event, they performed dataset cleansing, changing categorical features into numerical values and then performed feature scaling with different models. Through the use Oracle Analytics Cloud and Data Flows, the team demonstrated that the Support Vector Machine “SVM” provided the highest confident level of the different models tested. From there, further analysis was completed and a model was built with SVM in a notebook.



The Team:
Sheetal Gour, Salman Azhar, Nazzal Naseer, Muhammad Mustafa Ispahani, Odunayo Onifade, Noman Tanveer

The Problem / Challenge:
The team was focusing on the “Deep Fakes” challenge with the specific area of detecting whether the video contains a fake face given a URL of a video or a real target image.

The Solution:
The team decided to solve Deep Fake challenges by training the machine learning model against 3 different datasets: DeepFakes detection Kaggle, CELEB-DF, and the FaceForensics dataset. Over the weekend, the trained model on this datasets achieved an accuracy of 92%. By identifying what parts of the video are fake and placing boxes around them, it provides the confidence level of the model performance. This method enables the team to benefit from the advancements made in image segmentation with models having over 95% accuracies. They used Flask as the Frontend UI which calls an API that identifies the face in the video and labels it either real or fake.

TeamMaverick 2.0


The Team:
Siddhartha Bhattacharya, Nithya Subbaraman, Sam Sida, Rithika Venkatesan

The Problem / Challenge:
The team was focusing on the “Anomaly Detection” challenge with the specific area to understand more and visual the potential fraud events using the “Charlie’s Crash Repairs” story and dataset.

The Solution:
The team focused on finding a solution to address and identify potential solution to detect anomalies and prevent fraudulent transactions. The team developed a predictive model to predict the anomalies in the transaction feed. A dashboard was created to visualise the trends in the transactions and distribution of the amounts transacted; display the anomalous transactions with the likelihood of them being fraud and display the top suspicious accounts and visualisation of the inflow and outflow of these accounts.

anomaly detection 101010


The Team:
Shaunak Phaldessai, Andrenz Arfianto, Hirdesh Kumar, Sophanith Song

The Problem / Challenge:
The team was focusing on the “Anomaly Detection” challenge with the specific area being looking different purchasing patterns of large amounts.

The Solution:
The team built a solution that focused on high volumes of in and out transactions being made in a short period of time and deposits of large amounts of cash into accounts for purchasing assets eg property, cars, jewellery and gold.



The Team:
Raghu Kodanda, Vinay Chitrakathi, Cherishma Duggina, Vatsal Mavani, Usama Zafar, Gul Hassan, Shazia Rashid, KLN Suman

The Problem / Challenge:
The team was focusing on the “Anomaly Detection” challenge with the specific area being looking for cyclical transactions which are potentially fraudulent transactions.

The Solution:
The team wanted to provide a simple and user friendly solution to enable businesses and financial institutions, understand hidden patterns in their business data and better solve their problems. Another use-case is to enrich solution to incorporate more modelling techniques for financial frauds or data anomaly and deep fakes in the insurance domain.

Binary Bits


The Team:
Vaishnavi Borwankar, Catherine Chris, Cathy Jones, Tanaya Wadekar, Pranjal Sharma, Seyed Shaheen

The Problem / Challenge:
The team was focusing on the “Anomaly Detection” challenge with the specific area being around the following fraud scenario – Tony and Mary Trumpo are running a business with an expected turnover of $2.1M. Instead, it is operating with a turnover of $14.3M. The problem requires to identify where the additional $12.2M coming from.

The Solution:
The team studied the “Charlie’s Crash Repairs” datasets which were provided and analysed the data that fell under the high-risk zones. Jupyter Notebooks was used to figure out these anomalies and Oracle Analytics Cloud was used for the the data visualisation.

DeepFakes Busters


The Team:
Sujit Udhane, Ignacio, Kavish, Rishi, Hetaram, Ananya

The Problem / Challenge:
The team was focusing on the “Deep Fake” challenge with the specific area where digital assets such as ads based on audio, video and images can be easily manipulated by counterfeiters to generate fake ads and content.

The Solution:
The team proposed a solution which can detects the fake ad by applying Deep Fake AI models and also with technique of steganography (ie adding encrypted hidden message in original content) to prevent misinformation and fraud.



The Team:
Jia Wu, Mohamud Rashid, Hassaan Khhan, Asif, Praise Gombiro, Muhammad Usama, Muhammad Arham Adeel, Umer Nasir

The Problem / Challenge:
The team was focusing on the “Anomaly Detection” challenge with the specific area using the “Charlie’s Crash Repairs” story and dataset where they looked for fraudulent activities in 1.5 million transactions.

The Solution:
The team took a 4-step approach to compose an autonomous systematic anomaly detection. They evaluated transactions based on duplicated accounts, account risk levels, account categories and frequently referenced transactions.

Icode Cyber


The Team:
Uzma Qureshi, Mohammad Saad , Abdul Jabbar

The Problem / Challenge:
The team was focusing on the “Cyber Security” challenge with the specific area being database vulnerabilities.

The Solution:
The team used a variety of tools including Kali Linux, Burp Suite and Wireshark enabling them to focus on analysing these threats and then used Wix for hosting their web presence including a virtual assistant.

The Hackermen


The Team:
Muhammad Qasim Khan, Asad Imtiaz Malik, Fahad Zaheer, Muhammad Sharjeel Maqsood, Aatir Khan, Hamza Mushtaque, Mutahar Aamir, Arshanullah Tawhidi

The Problem / Challenge:
The team was focusing on the “Anomaly Detection” challenge where they hypothesised that people who have transferred an amount of > 200k of Australian dollars in a short period of time have a very high chance of being anomalous, or in simple terms, being engaged in fraudulent activities.

The Solution:
The team trained a K-Means Clustering Machine Learning model using Oracle Analytics Cloud and analysed the result using several visualization techniques including time series chart, bar graphs and network graphs. The analysis found that the hypothesis was correct after looking at the analytics dashboard.

This was just a sample of some of the teams. There are definitely others as there were over 100 teams submitting their projects and outcomes.

As mentioned, there are endless possibilities. If we are willing to learn, explore and collaborate.

#ThankYou #Tribute #CommunityMatters #ItTakesAVillage


Author: Jason Lowe

I am passionate about how organisations adopt IT quickly and sustainably to achieve a specific and measurable outcome. This thinking is supported through lean IT practices in operational support and project delivery, and adopting these practices with Oracle technologies by creating sustainable platforms. I'm also interested different perspectives and drivers - from enterprise to start-ups, academia to commercial to public sector, cost-sensitive (risk) to value-driven (reward) - all of which influences decisions that organisations make. I have a passion for community and have been called "a connector" - meeting new people that are trying to solve valuable and hard problems and connecting them with others that can validate and help realise their full potential. I've supported different organisations like TADHack and Hacking Health as a global organiser. I'm is a persistent blogger on medium.com and redthunder.blog and on LinkedIn - https://www.linkedin.com/in/lowe-jason #CommunityMatters #ItTakesAVillage

One thought on “#DigitalDefence – A Tribute To The Teams”

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: