EvalLens Newsroom author
Vladislav Starodubov
Co-founder, Engineering, EvalLens
Vladislav Starodubov is a co-founder of EvalLens and leads engineering. He builds the system behind it: judge orchestration, scoring infrastructure, security and repeatable evaluation runs. He writes about the evaluation pipeline, scoring consistency, calibration and bias in AI judges.
Writes about
- LLM evaluation pipelines
- Scoring reproducibility
- Model calibration
- LLM-as-a-judge bias
- Software architecture
Articles by Vladislav
ResearchSix lenses are not six models
ProductWhen every deck scores a five
Press ReleaseWe're building for the Build with Gemini XPRIZE
ResearchThe bias hiding in a single AI judge
ResearchSame deck, same score
ResearchFrom AI Jury to EvalLens: what 400+ runs taught us
Research