my perpetual goal is nextopt: ## Mission: affordable analytics for all - Reinforcement AI loop (user to user data to increased analytics of quality which pulls latent needs to induce greater number of users) cannot be initiated for small business people. Two reasons. First is the lack of infrastructure which prevents them from experiencing the benefits of analytics, second this ignorances which further makes them hesitant to invest in infrastructure. These problems needed a collaborative approach of business and small business people and further governments. ## Core competency 1. Predictive analysis capabilities with high accuracy and interpretability 2. Optimization capability to support decision-making by comprehensively considering demand forecasting results and user's operational status and satisfaction 3. Software capabilities that provide predictive analysis and decision-making optimization in the form of solutions Software capabilities that process large-scale and diverse data into usable information ## Company history 2016.08. Inventory optimization web service development 2016.12. Develop inventory management service for small bakeries 2017.05. Small business service Android app development 2017.08. Demand forecasting engine and web service development 2018.02. Nextopt corporation establishment 2018.06. LogisAll Demand Forecasting Contract Signed 2018.08. Selected as a Stan Korea sponsor company 2018.09. Developed Nextopt Demand Forecasting Engine ## Projects - grant project from the National IT Industry Promotion Agency ($85k) - Reliability prediction to optimize repair part inventory (battleship, airplane, tank), Client: Republic of Korea Armed Forces - Pallet Supply Circulation, Client: LogisAll (Pallet Rental Company) ($100k in total) - Exogenous Feature Engineering, Model selection, Approximate model (May-Aug 2020): accuracy improvement of 40% and 10 times faster speed - Self-developed Hierarchical Seasonal Model (Mar-Jun 2019): accuracy improvement of 30% - Auto-forecasting for Pallet Rental Company Using Stan Model: automated the entire 1 forecasting process with higher accuracy - e.g. ![[eg(nextopt).png]] ## product ![[Pasted image 20240101092734.png]] high resol: [[QTell 2.pdf]],[[QTell 3.pdf]] ![[Pasted image 20250606101111.png]]