Feb 2026 – Present
AI Specialist · Mindrift
- Automated web scraping and data extraction pipelines for AI training datasets
- Hybrid human/AI QA workflows for model output evaluation
- Cross-functional collaboration on data quality standards
AI data specialist at Mindrift. I build web-scraping and evaluation pipelines for AI training data, and on the side I'm shipping mytai, an AI fitness coach for iOS.
London · MSc Business Analytics at Bayes Business School
I work on the data side of AI: web scraping, dataset construction, and model evaluation. At Mindrift I build automated extraction pipelines and hybrid human/AI QA workflows for training data. Before that I spent a year at Scale AI benchmarking text-to-image models and evaluating conversational AI outputs for alignment and factual accuracy.
I'm finishing an MSc in Business Analytics at Bayes Business School, on top of a BSc in Information Management for Business from UCL. The dissertation is the last piece; after that, more time for building.
Outside work I'm shipping mytai, an AI fitness coach for iOS, and I run a small homelab. Away from a keyboard: football, powerlifting, and cycling around London.
Feb 2026 – Present
Nov 2024 – Nov 2025
May 2023 – Present
2025 – 2026
Bayes Business School
On track for distinction
Network Analytics · Deep Learning · Machine Learning · Revenue Management
2022 – 2025
UCL
Upper second class honours (2:1)
Languages & data
Python · R · SQL · Excel
AI & ML
RAG · LLM evaluation · Deep learning · BeautifulSoup · Selenium
Tools & infrastructure
Tableau · Docker · Ubuntu Server · Git
A fitness app with AI-powered video analysis for combat sports technique, food and exercise logging with barcode scanning, mobility drills, and a recovery dashboard pulling data from Apple Health.
React Native · Expo · TypeScript · Apple Health · iOS
Visit mytai.ukA personal home server on Ubuntu, running Dockerised services for smart home automation, network-wide ad blocking, and self-hosted apps.
Linux · Docker · Home Assistant · AdGuard Home · Networking
The quickest way to reach me is email: dragosmacsim@protonmail.com
I'm happy to talk about data and AI roles, mytai, or anything you've seen on this site. You can also find me on LinkedIn and GitHub.