About

About Me

🤖 I solve problems that require machine learning by deploying models to on-prem or cloud environments, monitoring the performance and scaling as needed.

☁️ When it comes to cloud platforms, I have a soft spot for GCP (not just because of the free $300 credits), but I'm also proficient in AWS and Azure .

💻 When I'm not busy cracking code, you'll find me dipping my toes into the open-source pool with projects such as African Whisper and DocIndex(I'm a maintainer of both), or taking the mic at Keras Community .

🏆 Oh, and remember that hackathon from Africa to Silicon Valley? Yeah, I strutted my stuff there too, landing a finalist finalist spot, another that I and a couple of friends won. So, you could say I thrive in the ring of challenges and competition!

🦀 If you want to talk about how Rust is better than all other languages, message me on Twitter.

Skills

Skills

Core Programming Languages: Python, Rust
Data Analysis & Visualization: Numpy, Pandas, Matplotlib
Machine Learning & Deep Learning: TensorFlow, JAX, PyTorch; expertise in Regression, Classification, Clustering, LLMs, LLVMs, Audio models etc
Cloud Platforms: Google Cloud Platform, Azure, AWS
MLOps and Monitoring: Docker, Kubernetes, Openshift, Prometheus, Grafana
CI/CD and Automation Tools: GitHub Actions, Jenkins, Terraform
Version Control: Git

Resume

Education

Data Science & Machine Learning

Massachusetts Institute of Technology(MIT)

Bachelor of Science & Computer Science

Catholic University of Eastern Africa

Professional Experience

Machine Learning Engineer

June 2024 - Present

SoCha LLC

  • Leveraged Machine Learning to analyze over 600 survey responses on perceptions of violent extremism in Northern Kenya and Somalia, significantly accelerating impact measurement for USAID’s counter-violence extremism efforts.
  • Optimized SQLAlchemy database schemas, enhancing data retrieval efficiency and performance.
  • Applied advanced NLP techniques to categorize survey responses, providing valuable insights on regional perceptions of violent extremism.
  • Developed a real-time dashboard to compare Language Model outputs with expert evaluations, streamlining decision-making.
  • Built a high-performance data ingestion pipeline, reducing processing time and improving data-driven decisions.
  • Visualized respondent sentiments using Pyplutchik research and open-source sentiment analysis models for more impactful insights.
  • Achieved over 80% test coverage (ongoing), ensuring high code reliability and robustness.

Contributor and maintainer

Mar 2024 - Present

African Whisper ASR Framework

  • Developed and maintained African Whisper, an open-source framework designed to enhance Automatic Speech Recognition (ASR) capabilities for African languages.
  • Implemented the methodologies from the following research papers:

    • Whisper-X: Time-Accurate Speech Transcription of Long-Form Audio for time-accurate speech recognition with word-level timestamps.

      Pyannote.audio: neural building blocks for speaker diarization for advanced speaker diarization capabilities.


  • Designed the framework to enable developers to fine-tune the Whisper model on multilingual audio datasets and monitor the training process.
  • Implemented features for developing and deploying a production API for audio file transcription or translation using Docker and Docker Compose for container orchestration.

Technical Consultant

Oct 2023 - Present

AlphaSights

  • I support AlphaSight's clients in exploring and implementing vector databases by comprehensively understanding their unique requirements, researching tailored solutions, and offering strategic guidance
  • Leveraging my expertise in vector databases, I enhance their data retrieval and LLM capabilities, providing valuable insights into the full potential of these databases.

MLOps Engineer

Dec 2023 - April 2024

Melian Dialogue

  • Contracted on a short-term basis with the UK research firm, executing the development of 10+ Minimum Viable Products (MVPs) showcasing the firm's research work across diverse AI applications.
  • Leveraged a tech stack that included Python and Rust languages, TensorFlow and PyTorch frameworks, as well as backend databases such as Postgres, and FastAPI for API development.
  • Contributed to substantiating and enhancing the impact of the firm's research initiatives by applying technical expertise and proficiency in AI technologies.
  • Managed cloud deployments and using Docker ensured optimal performance and scalability for the developed solutions.
  • Achieved 100% test coverage, ensuring robust and reliable code quality for all developed MVPs.

Machine Learning Engineer

Nov 2022 - Dec 2023

Freelance Contractor

  • Built an information retrieval application that empowers technicians to quickly identify error codes, root causes, and corrective actions for industrial PLC devices, minimizing downtime and improving operational efficiency.
  • Ensured application scalability and uptime by deploying and monitoring it on GCP.

Data Analyst Intern

June 2022 - Aug 2022

Kenyatta University Teaching, Referral & Research Hospital, Nairobi

  • Managed the complete workflow of retrieving, cleaning, analyzing, and presenting various datasets, incorporating clocking in data. Provided daily delivery and presentation of staff attendance information across multiple departments, contributing to technology rollout implementation and maintenance.
  • Showcased skills in data management, IT support, and collaborative project involvement. Acquired knowledge in handling diverse datasets, supporting IT infrastructure, and actively contributing to the successful execution of IT projects

Projects

My Projects


African Whisper

🚀 Seamlessly fine-tune and deploy Whisper model on multi-language datasets.


DocIndex

⚡️Fast persistent storage of multiple document embeddings and their metadata into Pinecone for RAG.


MemVectorDB

⚡️Lightning fast in-memory VectorDB written in rust🦀


Mail-QA

Mail QA is an AI workspace assistant that seamlessly integrates information retrieval, email management, and calendar scheduling into your workspace. This was part of a hackathon that got to the finals in Addis Ababa, Ethiopia.


MemVectorDB-Python-SDK

MemvectorDB python SDK.


Docker AWS CloudWatch Logging Application

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.


Deploy ChromaDB Instance on AWS using CloudFormation

Deploy ChromaDB on AWS CloudFormation using a template.


Automated Serverless-App Uptime Maintenance with Rust and GitHub-Actions

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.


MLOPs Pipeline for Plant Disease Image Classification using Pytorch.

MLOps Pipeline of an image classification model to classify images of plant leaves as healthy or diseased and the specific disease that a leaf is suffering from, if applicable, containerized using Docker and deployed on a Kubernetes cluster.


Medium-Article-Chatbot

Interact with one or many articles from Medium using conversational AI utilizing OpenAI GPT-3.5 model and get answers to queries based on the content of the Medium articles. Deployed as a REST API using Flask, containerized using Docker


GazeDirection-using-Mediapipe

Using the mediapipe and opencv libraries to track the Iris movement of a subject in real-time webcam video stream and calculate the gaze direction of the user.


Gesture Recognition using Mediapipe

Gesture recognition of 21 different gestures using Mediapipe from Google.


And Alot More in my Github

-->

Contact

Contact Me

Social Profiles

Email Me

keviinkibe@gmail.com

Loading
Your message has been sent. Thank you!