Developed an Assessment Engine and Recommendation Service using Python for personalized learning content extraction• Managed and optimized 100+ tables related to learning content, learners, and schools using PgAdmin and DBeaver• Optimized deployment pipeline for the Assessment Engine and eBook pipeline, shifting to Cloud Run for improved performance

• Developed an automated valuation model (AVM) for the Korean real estate market using XGBoost, achieving a better performance (MdAPE 3.3%) than Zillow AVM.• Implemented a data pipeline on Google Cloud Platform for model training and prediction, cleaning and geotagging 38 million data points from various sources.• Leveraged geographically weighted regression and clustering techniques to improve model performance.

Conducted a variety of the self-motivated urban & spatial research in the Bartlett Centre for Spatial Advanced AnalysisDeep Learning based Human Body Size Measurement Application (Mar 2022 - Sep 2022)Conducted a pro-retail computer vision based deep learning research to develop the anthropometry (body measurement) application with a higher accuracy. Using TensorFlow and OpenCV in Python, the app based on BodyPix and COCO(Common Objects in Context) achieved 95.84% of… Show more Conducted a variety of the self-motivated urban & spatial research in the Bartlett Centre for Spatial Advanced AnalysisDeep Learning based Human Body Size Measurement Application (Mar 2022 - Sep 2022)Conducted a pro-retail computer vision based deep learning research to develop the anthropometry (body measurement) application with a higher accuracy. Using TensorFlow and OpenCV in Python, the app based on BodyPix and COCO(Common Objects in Context) achieved 95.84% of accuracy in the experimentation of capturing the 4 body dimensions (arm length, leg length, chest girth and hip girth) for 10 volunteer subjects, which outperformed some of the previous DL research presenting 95.59% (Foysal et al., 2021) and 95.72% (Xiaohui et al., 2018). Flood AR - Am I at Risk? (Jan 2022 - May 2022)Built up an augmented reality web application that guides which flood risk zone of London the user is located in. Acquiring GPS location information from the user's smartphone, the app allows the user to interactively check the flood zone, flood alert and flood history of the locations. Aside from the AR application, I co-led the team, Delugeo, and created the webpage to raise the public's awareness the global and local flood vulnerability by visualising the related data. The project was acclaimed as one of the best projects in the programme and introduced as a sample work in the centre's promotion video. ▲ Project Website: https://nfabsikova.github.io/delugeo/▲ Web app: https://s-n-b-n.github.io/a/a.htmlLocation-Based AR Property Price Marker Application (Jan 2022 - Mar 2022)The application displays 2,366 price records of London Borough of Camden and Hackney sampled from 435,272 records in UK Land Registry's Price Paid Data. Using AR, the user can check the paid price for residential transactions via smartphone camera in real time. For geocoding, pgeocode package of Python was used. ▲ Price marker app: https://s-n-b-n.github.io/CASA0003/price.html Show less

Real Estate News Desk covering Housing, Real Estate Market and Urban Issues.

Founded a startup team. Made the minimal prototype of the data-driven education and course recommendation service, MNTR. Was able to successfully secure $100k angel investment. 2nd Winner awarded by TechStars.

Worked in RealtyGo, Real Estate Spinoff Media of the Chosunilbo, Korean largest news outlet.

Contributed to ‘Life Cycle Retail Marketing Segmentation Strategy Project’ (teamed with SAS Institute Inc.) of Woori Bank (KOSPI:000030) HQ in Seoul, Korea by collaborating with data scientists to extrapolate meaningful data that produced results for Woori Bank.

Executive Officer