BLUEPRINT GREEN DESIGN is an interactive, free, and lightweight app for anyone who wants to create and visualize simple building plans in 2D, calculate and visualize their heat loss, and get energy analysis for each component to make informed efficiency decisions.

A residential retrofit study for a single-family home in Puchheim, Bavaria. Combines envelope-driven heat-loss modeling, passive conditioning strategies, and RemRATE simulation to chart a staged path toward net-zero annual energy.

Hourly building energy-use forecasting on the ASHRAE Great Energy Predictor III dataset. Engineered temporal, and weather features and trained gradient-boosted models evaluated by RMSLE.

A platform concept addressing the communication and retention gap in the AEC industry, combining forums, job boards, firm reviews, and learning resources so practitioners and students can share knowledge across studios.

An explainable multimodal screening framework that infers building attributes from drone roof imagery, street-view facades, and electricity time series, generating retrofit action rankings for the a data-scarce neighborhood in Pittsburgh.

A transparent, multi-factor site-selection tool that scores all 3,234 US counties for data-center feasibility across 11 infrastructure, sustainability, and economic criteria. A web interface lets users set hard feasibility thresholds and soft sustainability weights, then explore the results on an interactive choropleth map overlaid with power plants and natural-gas pipelines.

Parametric daylight spaces to create two contrastive spatial experiences: one in an elegant restaurant and one in a banal warehouse, usingRhino Grasshopper and Climate Studio.

Transfer-learning classification of building exterior cladding materials from street-view façade imagery. Pre-trained ResNet50 and InceptionV3 backbones with frozen weights and a Dense / Dropout / Softmax head, trained on Wang et al. (2024)'s London subset and tested for geographic generalization on the Scotland subset.

Climate-aware energy and daylight simulation workflows for a civic space in Boston (ASHRAE 5A). Builds a reusable Grasshopper + Climate Studio template that ties wind-rose, humidity, and dry-bulb temperature analyses to sDA / UDI metrics.

3D urban building energy modeling from LiDAR point clouds in ArcGIS. Solar roof analysis via the Google Solar API, NREL PVWatts, and ArcGIS raster radiation tools to support retrofit and community-solar planning in Pittsburgh.

Interactive Power BI dashboard for the Larimer neighborhood. Geo-referenced timeseries electricity consumption and parcel property assessment data layers represented to inform the initial building stock screening.

Exploring the snow retention and building physics relations with off-the-shelf segmentation models and correlation analysis.

A weekend-scale climate / energy hackathon entry at MIT. Won the third place!

LEED v4 daylight and view analysis of a hospital using Climate Studio.

Parametric envelope and structural exploration using Grasshopper and Colibri.

Automating AIA 2030 export workflows mechanical engineers to accelerate the AIA2030 commitment. Exploring IESVE-Python API and to speed up comparative energy modeling in early design and batch submitting after design.

A low-cost greenhouse climate-control system based on LoRaWAN sensor networks and mobile-app alerts. Exploring autonomous response strategies to drifting temperature and humidity regimes using off-the-shelf distance, soil-moisture, and air-temperature sensors.

A speculative data-center–civic-space hybrid with an underground public amphitheater. Parametric daylight design in Grasshopper and Climate Studio paired with VR walkthroughs and glare overlays.

A student co-housing studio project exploring shared residential living.

A music school along Istanbul's Dolmabahçe waterfront. Integrates historical fragments and preserves green corridors with concert halls, dance halls, and rehearsal spaces opening onto a public conservatory campus.

Detailed construction documentation for a weekend house. REVIT 3D modeling, foundation/wall/floor connections, mezzanine acoustic design, and timber stud framing.

A home-empathizer concept that learns occupant patterns to optimize energy and comfort while prioritizing privacy. Built by a six-person interdisciplinary team (architecture, mechanical, electrical, business, industrial design).