Kevin Kibe's Projects
🚀 Framework for seamless fine-tuning of Whisper model on a multi-lingual dataset and deployment to prod.
This is a project I embarked on to determine what makes a song a hit by analyzing the audio features in Kenya from data scraped using the spotify API
This is a streamlit app for analyzing Mpesa Statements using Langchain Pandas Agent.
The objective here is to build an Artificial Neural Network that can look at Mel or MFCC spectrograms of audio files and classify them into 10 classes. The audio files are recordings of different speakers uttering a particular digit and the corresponding class to be predicted is the digit itself.
A script to keep an app I have on a free-tier serverless platform alive(not to spin down because of inactivity) by executing a GET request every 5 mins using the reqwests crate.
To build a predictive model that can find out the sales of each product at a particular store and then provide actionable recommendations
A book recommendation system using cosine similarity and tf-idf vectorization that outputs similar books to the user's input.
This is an application that enables a chat-based conversation in English with Swahili videos/audio files using Python and deployed as a REST API using Flask containerized using Docker.
A predictive model to identify customers who are at risk of attrition for a credit card service in order to find the key causes that drive attrition. This can help banks take appropriate actions to build better retention policies for customers.
This challenge is to predict the yield of rice crop for a given season at 100 geo locations (latitude and longitude) in the Chau Thanh, Thoai Son and Chau Phu regions of Vietnam using satellite data.
Using clustering techniques, to identify the several segments of customers to target the potential user base for a retail business.
The project aims to provide a solution for recruiters who struggle with determining appropriate salary ranges to offer candidates, as well as candidates who may be unsure about the salary ranges.
This project provides steps on how to deploy ChromaDB on AWS CloudFormation using a template.
Lets say there is a video game company being set up and they do not know what global region to market their products to. Using data from past video game sales this problem can be solved.
⚡️Framework for fast persistent storage of multiple document embeddings and metadata into Pinecone for production-level RAG.
This application facilitates running Docker containers and streaming their logs to AWS CloudWatch. It leverages Python's Docker SDK and Boto3 library for AWS interactions.
This is an application that allows users to interact with a text, pdf or word document using conversational AI techniques. The chatbot leverages the OpenAI GPT-3.5 model and Pinecone vectore store to deliver responses to user inquiries, utilizing the content of the loaded document.
a project that aims at creating a shorter version of a given text document while retaining the most important information.
dstack is an open-source engine for running GPU workloads. It simplifies development, training, and deployment of gen AI models on any cloud. Discord: https://discord.gg/u8SmfwPpMd
Email newsletter service written in Rust🦀.
This is an implementation of fine-tuning the Llama-2 model with the QLoRA (Quantized LoRA) framework using a specific version of Llama and a particular dataset all from HuggingFace Hub.
This repository features code for Swahili Automatic Speech Recognition (ASR) fine-tuning using the Whisper model on Swahili Common Voice dataset.