I'm a Data Engineer working on various optimization, sensoring, and analytics solutions to improve decision making prowess for stakeholders. Most of my projects are Python (flask, django, pandas), R (dplyr), Azure SQL, Azure Datalake, HTML, and Javascript based solutions with a ML or optimization engine. I believe in building small proof of concept projects, finding the correct value proposition and scaling them into large production systems.
Visit https://jasonbian.netlify.app/ to see more in-depth writeups on my projects.
I have an consulting company! Visit https://optimchain.github.io/Main/ to see how I can help you improve your supply chain operations.
Topics I'm interested in:
- Supply Chain Analytics
- Stochastic Processing/Modeling
- Supplier Classification and Regression Problems
- Time-Series Forecasting
- Edge Processing and Stream Analytics
- Cloud Event-Hubs and Telemetry Processing
- DTE Energy: Six Sigma ERP for contractors to centralize problem solving and data entry
- Cummins: Factory floor level re-order point models for tooling. Dynamic inventory management for bottleneck machines
- Microsoft: Regression models for datacenter capacity optimization, azure warehouse inventory buffering based on historical usage volatility, Azure data-lake based data engineering and dev-ops workflows
Heijunka Scheduler is a VBA and python based manufacturing scheduling solution with an point-in-time linear programming solve. Code is currently in production on two different product lines.
To see full implementation: Heijunka Whitepaper.pdf
Api for querying live league of legends data, concept here is to create personalized item analytics in R and python that give better decision insights than popular league analytics engines such as https://na.op.gg/
Repo: https://github.com/IamJasonBian/Riot-API-and-Match-History-Analytics
Cloud Based Sensoring Solution for flow telemetry.
Shown here is an early stage sensoring prototype with azure based alerting and app deployment. Code is currently proprietary and under development. Idea is to use Azure Cloud (Azure webapp, eventhub, azure datalake) to manage flow telemetry and provide an alerting/sensoring solution that is much cheaper than PowerBi/Stream Analytics solutions.
Repo: https://github.com/OptimChain/Cloud_Hydroponics
Full data science ready Van and Livable space. Combining Optimization and Van-Life concepts to create a sustainable life on the road.
Repo: https://github.com/IamJasonBian/Van_Life/blob/main/README.md
Cloud based ERP app for inventory management. Idea is to use QR code to automate the tracking of common standardized work elements that were previously done manual via excel.
Go to https://github.com/IamJasonBian/QR_ERP to see full solution.
3D Spatial Visualization contains 3d tools for visualizing time-series trajectories, go to the below link to see current live deployment:
https://sites.google.com/umich.edu/hfet-lab/lab-tools
https://github.com/IamJasonBian/Lab_Tools
Proof of Concept Machine Learning 3d Spatial Classification and Visualization via sensors:
Site: http://swingtech.io/
Repo: https://github.com/IamJasonBian/SwingTech_Core
Concrete Canoe is an University of Michigan sponsered project where optimal Hull/Hydrostatics canoe exteriors are created for concrete canoes via simulation. Exact parameters are located in the implementation below in the Hull Design section. Maxsurf simulations were used design of experiments style to solve for the optimal hull shape.
See Hull Design section for my contribution! This was the fastest canoe in the past 10 years in competition.
UniversityofMichigan_TECH PAPER.pdf
