I'm Kelvin (Bosire) Nyakundi

Software Engineer

Projects

Services

Image

About Me

Image

Hey everyone, I'm kelvin (aka Bos) and this is my web site - Boss Coders HQ. To give a brief introduction to myself, I am currently a Applied AI Software Engineer working at JP Morgan.

Ever since I was young I've always had a deep interest in computers especially coding, even though that was quite uncommon for the people I grew up around. This is often the case with others from urban communities where there is a lot of mystery around what it is that programmers do.

If it wasn't for the help of one teacher who spotted my interest early on and nurtured it, I would be in the same situation and would probably have gone a very different route

I hope that eventually, I can turn this personal website into a platform where I can pass knowledge on, so I can do the same thing that my teachers did for me and help other budding coders. If you'd like to stay tuned or support my cause in any way, please feel free to get in touch!

Contact Me

My Services

AI Engineer

Leverage the power of machine learning to automate processes, gain insights, and propel your business into the future. Tailored AI strategies and implementations that transform data into actionable intelligence.

Solutions Architect

Expertly crafted technology frameworks that streamline operations and enhance system efficiency. Strategic planning and technical design that ensures scalability and performance for your IT infrastructure.

Web App Development

Develop cutting-edge web applications that offer seamless user experiences with modern technologies and best practices. Custom web solutions that are as robust under the hood as they are beautiful on the display.

Graphic Design

Elevate your brand with compelling designs that capture attention and communicate your message effectively. From logos to layouts, we provide the visual assets that set the tone for your brand's identity and storytelling.

Projects

Ah I see you've made your way to the best section! Click on the buttons below too see some of the projects that I have worked on, I'm sure you'll find something thats interesting.


Mock AI - An AI Interviewer for Mock Software Engienering Interviews

Using LLM's & STT/TTS API's + Computer Vision

The vision is to develop an AI agent that delivers a mock interview experience on par with that of a human expert for software engineers. Mock.AI is an AI agent designed to deliver high-quality mock interviews for software engineers, matching a human expert's competence.

It will firstly focus on "no execution coding" type interviews, wherein candidates face a series of algorithmic and coding challenges. These interviews usually take place over Zoom, where the candidate communicates with the interviewer, explaining their code or pseudo-code solutions to problems in real-time, using platforms like a whiteboard or Google docs. This process aims to evaluate the candidate's coding proficiency, problem-solving, communication and coachability (i.e. ability to adapt to instructions).

GitHub


Fitnesso - Detecting Bad form

...Using Computer Vision and Machine Learning

Me and a team of developers are currently developing a cross platform mobile application using React Native that can detect bad form in videos of a person performing a gym exercise. The goal of this application is to combine computer vision techniques with machine learning models based off of convolutional neural networks to accurately analyse exercises. My team hasn't finished developing this yet but stay in tune for when we do!

Oxford Hack - AutoConfig

This was a project developed in a team at Oxfrod Hack 2019. We developed an Electron cross-platform desktop application that allowed users to quickly set up development environments using a click of a button. The final application used a combination of various new web technologies including React, Redux and Ansible's API. We plan to extend this to a mobile application as well as being able to set up environments using Amazon's Alexa so stay tuned!

GitHub

Hack the Midlands - Job Application Manager


This was a project that was developed independently at Hack the Midlands 2019. I Started the development of a Web-app Job Application manager that can store the state of a persons job applications and provide informative graphics based on statistics. The idea came to my head since a lot of my fellow final year students lost track of job applications while they were frantically looking for a graduate position xD. When I've finished development I will put the code up on my GitHub so we can all be more organised. Till then sit tight!

Virtual Trading Chatbot - AURORA


What was used: IBM Watson, Python, PHP, HTML, Materialize-css, MySQL

This was a project that was presented in collaboration with Deutsche Bank. It invovled creating a Chatbot using IBM Watson to provide answers to a user's queries on the stock market.

The key features of this project were:

  1. Access and interface with data on the FTSE 100.
  2. Receive and respond to queries from a trader. Queries can be in the form of voice or text commands.
  3. Employ AI so that it exhibits traits consistent with a personalised virtual trading assistant.
  4. Allow user to view news on any of the FTSE 100 without leaving the application.


Source Code Plagiarism Detection

...Using Machine Learning

What was used: Python, HTML, PHP, CSS, Scikit-learn, TensorFlow

This project involved me independently creating a Python software that used machine learning on large sets of data in order to detect source code plagiarism in Java files. To do this I utilised neural networks and natural language processing in order to detect 15% more plagiarism cases than commercial software.

GitHub
This was my 3rd year dissertation project which recieved a First Class. If you would like to see the full (100 page) report you can download it using this link.

Stars Classifier


What was used: Python, Scikit-learn, TensorFlow

This project came about due to a Kaggle competition that involved create the best classifier for astronomical data. I had to implement machine learning classification algorithms, including random forest classifiers, multi-layer perceptron's and convolutional neural networks.

Competition Link

Atari Breakout


What was used: C, OpenGl

Implemented a 3D version of the classic Atari Breakout game using only C and OpenGL. This was one of my favourite projects!

GitHub

Bill Splitter

Implemented an online bill splitter that had the following features:

  1. Provides support for groups of up to 10 people to split their bills.
  2. Keeps count of how much every person owes and is owned.
  3. Allows groups to see which bills have been paid, when bills are due and notifies them accordingly.


ProtoType Banking System

Created a prototype for a simple banking system which allowed for secure authentication and registration including 2-factor authentication and password recovery options. Involved producing a website as the front end and setting up a HTTP server as the backend, both of which were mainly coded using java to allow all operations to be done on the server side for security purposes.

Blog Posts

Unfortunately, there are no blog posts up yet, maybe check back in a couple of days and that will change!

Contact Form

The form is abit flaky right now sorry 😅 If you need to get a hold of me hit me up by email: Nyakundi@hotmail.co.uk