Editorial bio
Tudor Coman is a Senior Machine Learning Engineer at Adobe, seven years in, currently building what the industry still mostly talks about: a specialised team of AI agents that owns the full software development lifecycle – not a copilot bolted onto one phase, but automation across design, build, review and iteration.
His track record cuts through the usual AI-engineer profile: personalised recommendations, online adaptive experimentation via reinforcement learning, and now a large-scale A/B testing platform where AI agents generate and run customer-tailored generative tests. He also teaches AI-in-SDLC at the University of Bucharest, so the practice feeds the pedagogy and back.
Worth following if you want to see what agentic development looks like inside a company that already has the distribution to prove or kill it.
Key facts
- based in
- Romania
- current
- Senior Machine Learning Engineer, Adobe
- prior
- 7+ years at Adobe across fraud prevention, ranking/recommendation algorithms, LLMs
- teaching
- University of Bucharest – AI in software development & building apps with AI agents
- focus
- Agentic development · reinforcement learning · generative A/B testing
Featured work
Agentic SDLC team at Adobe
Building a specialised AI-agent team that handles design, build, review and iteration across the software development lifecycle.
Large-scale experimentation platform at Adobe
AI-agent-driven generation and running of customer-tailored generative A/B tests.
University of Bucharest (Instructor)
Course on AI in software development and building applications with AI agents.