Leo Milojevic
ML Systems Engineer | Contract and advisory engagements
What I do
My work covers RAG and LLM systems, multi-objective optimization, and pricing — across consumer ML products serving tens of millions of users and retail pricing for 1,000+ stores. The work I'm best at: designing new ML systems that have to actually work in production, integrating ML into systems with existing logic rather than replacing them, and models that survive contact with messy real-world data.
Recent work
Consumer edtech · 2025–present
Working on production RAG/LLM systems for a large US education platform, focused on retrieval relevance, evaluation infrastructure, and scalable knowledge-centric ML workflows. Specific implementation details remain under NDA.
Studenac Market · 2022–2025
Built and iterated on machine learning pricing systems for a large Croatian retailer, working closely with pricing leadership and McKinsey on price optimization across diverse retail environments and product groups.
Collaborated with university researchers on modeling the impact of Croatia’s retail Sunday trading restrictions, focusing on store-opening strategy under regulatory and competitive constraints.
Retail trading system optimization · 2021–2022
Researched systematic trading strategies using statistical modeling and simulation in Python, contributing to internal tooling for backtesting, stress-testing, and evaluating behavior across changing market conditions.
Writing
Working together
I take on contract engagements ranging from short focused diagnostics to ongoing implementation work. If you're working on something where senior ML help would be useful, the easiest way to start a conversation is by email.