I Putu Cikal Taruna

• Built Drillo.id from zero to launch: a self-learning drilling platform with structured practice sets, explanations, instant feedback, learning materials, and live tryouts for learners from elementary school to professionals

• Designed the initial product experience and runs product management end to end, from roadmap and prioritization to delivery with engineering

• Set up the entire content foundation, including question banks, explanations, and learning materials across all learner segments

• Directs AI product strategy and go-to-market execution for AI-native assessment platform with LLM-powered question generation, live scoring, and real-time ranking

• Leads technical architecture while remaining hands-on with engineering, building production AI systems using Python and Golang

• Partners with product and engineering leads to define requirements, prioritize features, and manage delivery milestones

• Builds strategic partnerships and investor relationships to secure funding and accelerate company growth

• Scaled AI engineering for platform serving 200K+ users, managing end-to-end ML deployment and model evaluation

• Architected RAG pipelines generating 5K+ questions monthly, reducing content creation time by 60%

• Improved output accuracy by 30% and reduced AI processing costs by 50% through model evaluation and LangFuse token optimization

• Evaluated GPT-4, Gemini, and Perplexity capabilities to establish model selection framework across use cases

• Directed 20+ person cross-functional team spanning C-Level, Academic, Marketing, and IT for national competition serving 8K+ students

• Built data systems and workflows enabling real-time simulation tests at national scale

• Managed national competition registration with PUSPRESNAS and SIMT, coordinating government compliance documentation

• Led 3-member QA team across 10+ applications serving 200K+ users, reducing post-release defects by 25%

• Built automation framework with Python, Selenium, and Grafana K6, cutting testing costs by 40% and boosting efficiency by 50%

• Achieved 30% decrease in bug resolution time through structured documentation and streamlined release cycles by 20%

• Improved AI-generated content quality by 50% building generation tools using Python, LangChain, GPT-4, and Gemini

• Reduced AI processing costs by 50% implementing LangFuse token efficiency strategies across production systems

• Improved content accuracy by 30% through user feedback analysis and optimized prompting techniques

• Led 7-member cross-functional team building B2B exam platform using Agile methodology with Notion and GitHub

• Boosted exam creation efficiency by 80% designing AI-driven content generation and question randomization tools

• Developed Business Requirement Documents (BRD) aligning system design with business needs and stakeholder expectations

• Increased student retention by 25% building executive dashboards using Looker Studio and PostgreSQL

• Improved decision-making by 35% creating interactive dashboards with real-time student performance and business metrics

• Improved prediction accuracy by 20% developing predictive models with Python to forecast student exam outcomes

• Reduced manual workloads by 40% automating data reporting tasks using Python and SQL

• Built core SmartBTW mobile app features serving 200K+ users with Flutter, Dart, and Firebase, including fitness tracking, real-time location mapping, and psychometric testing modules

• Implemented Samapta sports calculation feature with GPS tracking, expanding student assessment from academic to physical fitness metrics

• Shipped mobile-first features in collaboration with cross-functional product, design, and backend teams for Indonesia's leading EdTech platform