Damir Olejar

Provided strategic technical consultation on AI feasibility, architecture, and integration for broadcast and newsroom products, advising stakeholders on computer vision, audio understanding, and NLP solutions. Designed and delivered an AI-driven customer support chatbot using LLMs, RAG, and semantic search, enabling autonomous troubleshooting and reducing support escalations. Built AI-assisted newsroom tools including a multimodal video search system (Metamorph) and an automated script-writing assistant (Scriptor), demonstrating end-to-end AI application development from concept to production. Collaborated with product owners and journalists to define requirements, scope timelines, and communicate technical trade-offs, ensuring alignment across engineering and non-technical teams.

Led modernization of a flagship AI sales analytics platform, replacing legacy code with Python/Java microservices and RESTful APIs serving 100,000+ users. Directed a cross-functional team of 10 engineers through full SDLC using Agile/Scrum, managing project scoping, sprint planning, and delivery roadmaps; reduced production bugs by 50%+. Engineered NLP and machine learning pipelines using LLaMA 2 and BERT, improving opportunity prediction accuracy from approximately 75% to approximately 90%. Served as key technical liaison to business stakeholders, translating architectural decisions and NLP capabilities into actionable product requirements and business impact.

Designed and validated an NLP prototype for health data using Apache Spark, Java, spaCy, and LangChain, delivering 30% additional functionality to internal data science platforms. Partnered with Product and Research leads to align technical execution with roadmaps, accelerating project delivery by 20% through clear requirements translation and cross-functional workflow optimization. Mentored junior developers and established engineering best practices, strengthening team throughput and code review processes.