Frimpong Richard

Experience

Language & Tools Experience

• Languages: Java, Typescript, Python, PHP, JavaScript
• Frameworks: Springboot, Laravel, PHP Slim, Vue, Angular
• Databases: MySQL, Postgres, Redis, Mongodb, OrientDB
• Tools: Docker, Jenkins, AWS, Harshicorp Vault, Ansible, Vagrant, RabbitMQ, Keycloak

Pet/Hobby Projects

• Find from my followers who (Twitter Project) GitHub Repository
  This project is to find my followers who can help me with a request based on their profile picture, tweets, replies, bio/description and likes
  For example, if I want to buy a car, I can find a follower who is a car dealer or has a car dealership based on some criteria about the follower's information.
  
  The algorithm runs as follows:
  1. Profile picture: If the profile picture is car or automobile-related, then the follower is a potential car dealer
  2. Tweets: If the follower tweets about cars, then the follower is a potential car dealer
  3. Replies: If the follower replies to tweets about cars, then the follower is a potential car dealer
  4. Bio/Description: If the follower's bio/description contains car dealer, then the follower is a potential car dealer
  5. Likes: If the follower likes tweets about cars, then the follower is a potential car dealer
  The algorithm then ranks the followers based on the number of criteria met and gives the ranking in chronological order
  
  Tools:
  Integration: Twitter APIs, Google's search APIs
  Backend: Java Springboot
  Framework: Apache Camel
  Hosting: AWS
  AWS Services used: EC2, S3, Lambda, API Gateway, Cloudwatch, SNS, SQS, Route 53, VPC, IAM, Cloudfront, Cloudformation, Rekognition, CodeDeploy, CodePipeline
  CI/CD Tool used: Jenkins
  Enterprise Integration Patterns Used: Splitter, Aggregator, Content-based router, Message filter
  Databases: Redis
  Message broker: RabbitMQ
  





• Rent Prediction With Linear Regression GitHub Repository
  This is a machine learning project that uses linear regression to predict rental prices based on various factors such as location, square footage, number of bedrooms, and amenities.
  The model is trained on historical rental data to identify patterns and relationships between features and rent prices.
  By leveraging regression analysis, this project helps in estimating fair rental values, assisting landlords, tenants, and real estate professionals in making data-driven decisions.
  Tools:
  Language: Python
  Libraries: Numpy, Pandas, Seaborn, Matplotlib
  
• 
  
  
  
  Churn Prediction With Logistic Regression GitHub Repository
  
  This is a machine learning project that aims to predict whether a customer is likely to leave a bank (churn) based on factors such as account balance, transaction history, customer tenure, and service usage.
  By applying logistic regression, the model estimates the probability of churn, helping banks identify at-risk customers and take proactive measures to improve retention.
  Tools:
  Language: Python
  Framework: Scikit-learn, NumPy, Pandas
  

Some Security Findings

In 2017, I found out that Lead for building OMGVoice's website website was vulnerable to XSS.
They redesigned the site shortly after I informed them and the new one doesn't have such vulnerability.

In 2022, I discovered two security vulnerabilities on DVLA, i.e., the Driver and Vehicle License Authority in Ghana's official site.
The site used users' phone numbers as usernames and passwords for users who did not change their default password after the first time logging in, which is probably 99% of users.
This was a big concern, as once entered, you aren't forced to reset your password; hence, you could access anyone's account if you know their phone number (both at the UI and API level).
The problem this poses is that a bad actor can write a script to randomly generate giant records of phone numbers and then use the API to authenticate the generated phone numbers.
If authentication is successful for a phone number, the person can then obtain other PII such as DOBs, emails, identification card details, etc.

The second vulnerability was excessive information exposure. The site also had information such as fingerprint information (that is the format that fingerprint information is stored in) exposed on the API but not displayed on the UI.
This also poses a significant security problem.
I informed DVLA of my findings, and they have since implemented an OTP that is sent to the user's phone for confirmation as part of the authentication process for the first problem.
They have taken out the fingerprint information that was returned in the API response.

There are other vulnerabilities I have found in other applications in Fintech apps and even Telecom applications.
Unfortunately I can't post them here as they still exist and can be exploited.
If they are resolved, I will post them here.
But the key takeway is that please take application security very serious.

Conference Talks & Publications

• Road traffic identification and measurement system using object recognition by Frimpong Richard and Matey Aryeh


• A talk on AI And Financial Inclusion at the Ghana-India Kofi Annan Centre of Excellence in ICT (GI-KACE) in August 2022 by Frimpong Richard & others.
  (My talk starts from 34:30 to 46:08 in the video)


• An article by Frimpong Richard to the Government of Ghana on the need for financial inclusion and artificial intelligence.

Education

Interests

Apart from being a software developer, I enjoy most of my time being indoors.
I am an avid fan of history especially world war histories.
I also delight in reading a lot about military strategies and technological innovation as well as business and financial articles

In Progress

To force myself to release early and often, I have listed some of my projects/articles that are in progress in their draft form.
The rational is to experience the shame in having an unfinished project or article in the open and to force myself to complete it to avoid such shame 🤣🤣🤣🤣