How AI/ML shapes EPAM into the Adaptive Enterprise
summary
EPAM builds a company-wide skills graph and cloud-native event pipeline that converts CVs, course syllabi and event signals into machine-readable skill objects for ~40,000 people. Key components include text-to-skill model, skill MDM, a Skill Hunter extractor, semantic search and a primary-seniority prediction model; ML outputs use multi-event trust/confidence scoring and human-in-the-loop assessment committees. Architecture emphasizes graph DBs, predictive analytics for staffing and internal mobility, and reuse of pre-trained NLU (eg BERT/GPT) where it speeds development.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 1:04Skills in modern enterprise
- 4:36Skills ecosystem at EPAM
- 11:59Skills platform overview
- 23:21Non-linear competency structures
- 27:57Skills system capabilities
- 34:18Humans in adaptive enterprise
- 36:28Q&A on skill accuracy
related talks
06mIntroducing EPAM
30mis designing ML solutions to deliver human capital insights
34mOrganized chaos adopting GenAI at Enterprise
22mAgentic AI for Engineering Leaders | Strategy & Execution Guide
01h 00mEnterprise Architecture for Next Gen Innovations
47mEngineering Management in the AI Era
16mEngineering the Future with the Reliable Delivery Management
31mBuilding Agentic AI Platforms