Moses B.

When you hire a developer, what you are really looking for is someone you can trust to take your project seriously and actually get it done. That is what I try to be for the people I work with. Over the past four years I have built entire products: the server and database running behind the scenes, the website or app people actually use, the automation that handles the boring work, and the AI features on top. I have done this for clients across several continents, both on and off Upwork. I work mainly in Python and JavaScript, with FastAPI, Svelte, AWS, Docker and more. SKILLS Full-Stack Development · Python · JavaScript · TypeScript · FastAPI · Flask · Angular · Svelte · AWS · Docker · REST API · PostgreSQL · MongoDB · Web Scraping · Selenium · Automation · Machine Learning · Computer Vision · Generative AI · OpenAI API · WebRTC. Clients keep working with me because I am reliable, I communicate clearly, and I do not disappear in the middle of a project. My reviews and portfolio below show the actual work, so you do not have to take my word for it. If you have a project in mind, send me a message with a few details and I will tell you honestly whether I am the right fit. What I have worked with: ✅ Web apps and websites: Angular (with Angular Material and PrimeNG), JavaScript, TypeScript, HTML, CSS, Firebase ✅ Backend and APIs: FastAPI, Flask, REST APIs, WebSockets, gRPC, and high-performance async Python (asyncio, threading, multiprocessing) ✅ Automation and web scraping: Python and Selenium ✅ AI and machine learning: PyTorch, NumPy, Pandas, data gathering and preprocessing ✅ Natural language processing: Transformers, Gensim, and vectorizers like TF-IDF ✅ Audio and speech: Whisper, ffmpeg, librosa, soundfile, PulseAudio ✅ Databases: PostgreSQL, MySQL, SQLite, MongoDB ✅ Cloud and DevOps: AWS (ECS Fargate, EC2, S3, Lambda, API Gateway, ECR, CloudWatch, EventBridge), Docker, Git and GitHub, and Bash scripting. PORTFOLIO These are a few of my open-source projects. You can read the full code for them, and browse more of what I have built, on my GitHub account. Comic panel detector (computer vision) An open-source machine learning model I fine-tuned to automatically find and separate the individual panels on a comic book page. Handy for anyone digitizing comics, building a reader app, or feeding pages into a larger pipeline. My most-liked project on hugging face. Built with: Python, YOLOv12, PyTorch. Driver drowsiness detection A vision system that watches a face and flags signs of drowsiness as they happen, the kind of thing that goes into driver-safety or attention-monitoring products. I built two versions to compare approaches, one on YOLOv11 and one on MobileViT. My most upvoted project on Kaggle. Built with: Python, PyTorch, YOLOv11, MobileViT. FAQ chatbot that answers from your own documents A fast retrieval-based (RAG) chatbot that answers questions from a specific set of FAQs or company documents instead of making things up. This is the pattern behind most genuinely useful business chatbots today. Built with: JavaScript, natural language processing, retrieval search. AI article generator A tool that turns a short prompt into long-form text like articles and blog posts using OpenAI's models. A simple, practical example of putting generative AI to real work. Built with: Python, OpenAI API. Yoruba dictionary web app A clean, fast dictionary website holding more than 8,000 English to Yoruba words and phrases, with an African-inspired design and search that helps people find what they mean, not just what they type. A full web app, designed and built end to end. Built with: Angular, TypeScript, Python. Zoom meeting bot Give it a Zoom link and it joins the meeting on its own, using browser automation that behaves like a real person rather than a script. A good example of the repetitive work I can take off your hands. Built with: Python, Selenium, undetected-chromedriver. Thank you for reading through my profile and best regards. I look forward to hearing from you.