Software Engineer & Machine Learning Enthusiast

Software engineer specializing in ML-powered APIs and backend systems with hands-on delivery using Python, FastAPI, Docker, and cloud deployment platforms.

Quick Highlights

  • Led a 4-member Agile team to build and deploy a real-time toxicity & profanity detection API.
  • Delivered 4 cloud-based subsystems (Detoxify model + Safe-Text API) with Dockerized services.
  • FoodVision classifier achieved 91% training and 93% validation accuracy.

About

I’m Moges Tesema, a software engineer and ML enthusiast focused on building reliable APIs, deploying models to production, and shipping data-driven systems. I’ve led a small Agile team to deliver a real-time toxicity detection API and built ML models ranging from CNN-based classifiers to regression systems.

Professional Experience

Machine Learning Engineer | Scrum Master, Intern

Ethronics Institute of Robotics and Automation Systems · Adama, Ethiopia · 06/2025 – 09/2025

  • Led a 4-member Agile team to build a real-time toxicity & profanity detection API using microservices.
  • Optimized model performance with quantization for faster inference and lower memory usage.
  • Coordinated Jira workflows and deployed 4 cloud-based subsystems, including the Detoxify model and Safe-Text API.
  • Stack: Python, FastAPI, Flask, PostgreSQL, Docker, Render, Neon, Railway, Hugging Face.

Core Skills

Backend & APIs

API design, microservices, and deployment for ML systems.

FastAPI Flask REST APIs Microservices

Machine Learning

Model training, evaluation, and deployment workflows.

TensorFlow Keras PyTorch Hugging Face

Data & Deployment

Dockerized services, cloud hosting, and analytics tooling.

Docker Render Railway Pandas

Languages & Tools

Primary languages and MLOps tooling.

Python JavaScript Java MySQL Git Weights & Biases TensorBoard Kaggle

Featured Projects

Safe-Text API Service

Live API

Real-time toxicity and profanity detection API built with a microservice architecture. Delivered as part of a 4-member team with cloud deployment.

FastAPI TensorFlow Docker

Model Demo

FoodVision (CNN + EfficientNetB0)

GitHub

Fine-tuned EfficientNetB0 on Food-101 and achieved 91% training and 93% validation accuracy using data augmentation and transfer learning.

TensorFlow CNN Transfer Learning

House Price Prediction (Neural Regression)

GitHub

Built a neural network regression model on Kaggle data with normalization and hyperparameter tuning for accurate predictions.

Neural Networks Regression Kaggle

Complete Python Bootcamp Projects

Repo

Advanced Python exercises covering OOP, concurrency, generators, decorators, and ML pipelines.

Python OOP Concurrency

LeetCode DSA Solutions Hub

Repo

Organized DSA solutions for LeetCode and Codeforces challenges.

DSA Algorithms Optimization

Education

B.Sc. in Software Engineering

Adama Science and Technology University · Adama, Ethiopia · 02/2023 – Present

Relevant coursework: Data Structures & Algorithms, OOP, Operating Systems, Computer Networking, Database Systems, Security, Software Architecture & Design.

Competitive Programming Community Education

Africa to Silicon Valley · 03/2025 – 09/2025

Focused on graph & tree algorithms and dynamic programming.

Let’s Connect

Open to software engineering and ML roles, internships, and collaborations.