Hi there, I'm Michael Moses! 👋

I'm a Mathematics & Computer Science student, AI Engineer, and Backend Developer passionate about building intelligent software and turning AI research into practical products.

I work across the entire AI engineering stack — from data, embeddings, RAG pipelines, and LLMs to backend APIs, databases, deployment, and production infrastructure.

I'm particularly interested in making AI systems smaller, faster, cheaper, and more useful in real-world applications.

• 🤖 Building Generative AI, LLM, RAG, and AI agent systems
• 🧠 Working with LLM fine-tuning, LoRA, distillation, quantization, and evaluation
• 🔎 Building semantic search and hybrid retrieval systems
• ⚙️ Developing production Python and TypeScript backends
• 🏗️ Building AI-powered products from idea → architecture → deployment
• 🐳 Working with Docker, CI/CD, Linux, VPS infrastructure, and cloud deployments
• 📱 Building cross-platform applications with React Native / Expo
• 🎓 Studying Mathematics & Computer Science at JKUAT
• 💻 Currently working primarily with Python, TypeScript, JavaScript, and Java
• 🌱 Always experimenting with new models, frameworks, and AI engineering techniques

🔭 Current Focus

• LLM & Generative AI Engineering
• RAG & Hybrid Search
• AI Agents & Tool-Using Systems
• LLM Fine-Tuning & Distillation
• Model Quantization & Efficient AI
• Backend & API Engineering
• AI Product Development
• Deployment & AI Infrastructure

🧠 AI & Machine Learning

• Large Language Models (LLMs)
• Generative AI
• Retrieval-Augmented Generation (RAG)
• Semantic Search
• Hybrid Search
• Vector Databases
• Embeddings
• AI Agents
• Prompt Engineering
• LLM Fine-Tuning
• LoRA / PEFT
• Knowledge Distillation
• Model Quantization
• QAT
• Model Evaluation
• NLP
• Machine Learning

🤖 Models & AI Ecosystem

• Qwen / Qwen-Coder
• Llama
• DeepSeek
• Phi
• Claude
• Sentence Transformers
• Hugging Face
• llama.cpp

⚙️ Backend Engineering

• Python
• Django
• Flask
• FastAPI
• TypeScript
• NestJS
• Node.js
• REST APIs
• Authentication & Authorization
• Background Jobs
• Webhooks
• API Integrations
• Database Design
• Microservice Architecture

🗄️ Databases & Data

• PostgreSQL
• pgvector
• Supabase
• MongoDB
• SQLite
• Firebase
• FAISS
• Vector Search
• Full-Text Search
• Database Optimization

🌐 Full-Stack Development

Frontend

• React
• React Native
• Expo
• JavaScript
• TypeScript
• HTML
• CSS
• Tailwind CSS
• Bootstrap
• Vite

Backend

• Python
• Django
• Flask
• FastAPI
• Node.js
• NestJS

🛠️ DevOps & Infrastructure

• Docker
• Docker Compose
• Git
• GitHub
• GitHub Actions
• Linux
• VPS Deployment
• CI/CD
• Nginx
• Cloud Deployment
• Containerized AI Services

🔧 Tools & Technologies

🧠 Semestra

An AI-powered academic platform designed to become a digital academic operating system for university students.

Built around:

• AI-powered learning
• RAG-based academic assistant
• Semantic and hybrid search
• AI-generated flashcards
• Notes management
• Past papers
• Class management
• Realtime collaboration
• Subscription infrastructure

Stack: React Native, Expo, Supabase, NestJS, PostgreSQL, pgvector, LLMs

🤖 CodeMate-Qwen

An experiment in building specialized coding models through fine-tuning, distillation, merging, and quantization.

Worked with:

• Qwen Coder models
• LoRA / PEFT
• Supervised fine-tuning
• Knowledge distillation
• GGUF conversion
• llama.cpp
• Model evaluation

The goal is to build capable coding assistants that can run with smaller computational requirements.

💻 Mkuu Code

A VS Code AI coding assistant built around an agentic coding engine.

The project explores:

• AI coding agents
• Tool calling
• Codebase understanding
• Agent orchestration
• VS Code extension development
• Local AI tooling
• AI-assisted software engineering

🔎 AI Search & RAG Systems

I've worked on retrieval systems combining:

• Vector similarity search
• PostgreSQL full-text search
• pgvector
• Embeddings
• Reciprocal Rank Fusion
• Reranking
• Semantic retrieval

The goal is to build RAG systems that retrieve the right information, rather than simply retrieving the most similar text.

💳 Financial AI & Automation

I've also worked on AI systems for financial applications, including:

• Transaction analysis
• Fraud detection concepts
• Automated classification
• Payment API integrations
• Customer-support AI
• AI-powered developer assistants

📚 Currently Learning & Exploring

• Advanced LLM architectures
• Agentic AI
• Efficient model training
• Quantization-Aware Training
• Small Language Models
• AI infrastructure
• Distributed systems
• Advanced backend architecture
• Mathematical foundations of Machine Learning

🎓 Education

BSc Mathematics & Computer Science

Jomo Kenyatta University of Agriculture and Technology (JKUAT)

My background in mathematics gives me a strong foundation for understanding the theory behind:

Linear Algebra → Probability → Statistics → Optimization → Machine Learning → Deep Learning → AI

📊 GitHub Stats

I'm always interested in:

• 🤖 AI & LLM projects
• 🧠 Machine Learning research
• ⚙️ Backend engineering
• 🚀 AI startups
• 🛠️ Open-source projects
• 💡 Building useful products

Portfolio: https://michael-moses.onrender.com/

Email: mosesmichael878@gmail.com