Emmanuel Onwuegbusi

Software Engineer with experience building production AI systems.

I build reliable software across backend systems and AI-powered products.

dev.to

About Me

I engineer production software for AI-heavy products, from backend architecture to end-user interfaces.

I focus on robust APIs, scalable data systems, and practical AI workflows that solve real operational problems.

Core focus: reliable backend and AI systems.

Skills

Agentic AI Systems Multi-Agent Workflows RAG Pipelines Structured Outputs Evaluation & Reliability Tool Calling / MCP Patterns Python & FastAPI TypeScript & React PostgreSQL Docker & AWS CI/CD & Automated Testing

Featured Projects

psqlomni

Natural language SQL assistant for real databases

Transforms plain English prompts into SQL with human approval before execution. Uses LangGraph orchestration with retry loops and supports OpenAI, Anthropic, Gemini, and Ollama providers.

Stack: Python, LangGraph, LangChain, SQLAlchemy, PostgreSQL, CLI tooling, automated tests.

View repository

psqlomni demo gif

screenshotToCSV

AI vision extraction app for structured data workflows

Built a full-stack extraction platform where users upload screenshots, define schema fields, and export normalized CSV rows. Includes BYOK model access, multi-image processing, and provider-aware error handling.

Stack: FastAPI, Next.js, TypeScript, OpenAI Vision, Gemini, CSV processing, production logging.

View repository

Agentic Framework

Open-source framework for production AI agents

Agentic is an opinionated open-source framework for building AI agents that use tools, collaborate in teams, and run in production with API and UI interfaces.

Its wow factor is that it turns agent ideas into reliable systems with orchestration, event-driven execution, human-in-the-loop control, and deployable runtime patterns.

Stack: Python, FastAPI, Ray runtime, LiteLLM, RAG integrations, Next.js and Streamlit interfaces.

View repository

AI Support Workflow Automation

Agentic workflow for intelligent Gmail support operations

AI support workflow automation diagram

This workflow monitors incoming Gmail messages, classifies intent, and routes support emails into an AI agent path.

On the support path, an agent uses memory, OpenAI chat models, and Pinecone vector retrieval to generate context-aware responses, then labels emails and creates drafts automatically.

In parallel, a Google Drive ingestion flow downloads files, chunks them, embeds them, and updates Pinecone so the agent always answers from fresh knowledge.

Stack: OpenAI chat + embeddings, Pinecone vector store, Gmail/Google Drive integrations, text splitting, automated workflow orchestration.

Previous Projects

OCR and Summarize Web App

A web application text extracts text from an image and summarizes the text for you.

Click to check out the app

Article    Code


Toxic Comment Detector API

An API that detects if a comment is toxic or not.

Article    Code


Malaria Parasite Specie Classifier

Classify malaria parasites by specie.

Code


Achievements

Winner Sea Turtle Rescue: Error Detection Challenge!

Built system to help Kenyan non-profit organization "Local Ocean Conservation" identify potential errors and anomalies in their sea turtle rescue database.

Click for more details

Winner AI Art Challenge on Zindi

Produced master music piece leveraging Artificial Intelligence.

Click for more details

Blog Posts

How to Run Llama 3.1 Locally in Python Using Ollama + LangChain

Step-by-step guide to running Llama 3.1 on your local machine and wiring it into Python applications with Ollama and LangChain.

Click for more details

Create Your Own Custom LLM Agent Using Open-Source Models (Llama 3.1)

Build a custom AI agent with open-source LLMs, including agent structure, tools integration, and practical setup patterns.

Click for more details

Build Your Own AI Agent That Can Browse the Web and Take Actions

Shows how to design an agent that can navigate websites, gather information, and execute actions in browser-driven workflows.

Click for more details

How to Build a Talk-to-Your-Data Chatbot Using OpenAI, LangChain, and Streamlit

Build a chatbot that answers questions over your own data using OpenAI models, LangChain orchestration, and a Streamlit interface.

Click for more details

Build a Chatbot to Interact with Your Pandas DataFrame Using Reflex

Explains how to create a chatbot interface for DataFrame exploration with Python, Pandas, and Reflex-based UI patterns.

Click for more details

Build an OCR and SUMMARIZE Webapp using pytesseract, gensim, and Django

In this tutorial, we are going to build a web app that extracts text from an image and summarizes it with gensim.

Click for more details

Testing in Software Development

In this article, you will learn what software testing is, why it matters, and how to implement a unit test.

Click for more details

Build a Todo API backend using Generic class-based views in Django Rest Framework

Build a to-do API backend using Django Rest Framework with create, read, update, and delete support.

Click for more details

How to use mixins, Viewsets & Routers to create a Todo API in Django Rest Framework

How to build the same Todo API using mixins, viewsets, and routers.

Click for more details

Develop an e-Commerce Store using Django for Beginners

Develop an e-commerce store where users can add items to a cart and make payment.

Click for more details

How to Contribute to Open Source Project

Learn to contribute to open-source projects on the Internet.

Click for more details

Contact

Let’s build reliable software and AI products together.

Email:     emmamichael65 at gmail dot com