from 0 to 29
Hillside reaches with the ground/water split shown (of 34)
A browser twin that shows where a release travels over real terrain, how much reaches water and how certain each result is.
- JavaScript
- Node.js
- CesiumJS
- 01 · Case studyA verifiable 3D spill-simulation twin for pipelines and channelsfrom 0 to 29Hillside reaches with the ground/water split shown (of 34)
- 02 · Case studyA wear and remaining-life digital twin for a slurry pipelinefrom 1 to 7Flow regimes the wear law tells apart
- 03 · Case studyGPU-aware orchestration of live 3D maps in an operations dashboardfrom 75% to 0%Dashboard maps rebuilt after an 11-minute absence
- Projects22
- Case studies17
- Layers6
- Tools121
01Work
Built end to end.
Filter
22 projects · case studies first
from 1 to 7Flow regimes the wear law tells apart
Charges every operating excess to the pipe in days of life, per segment and per minute; built by a 30-agent AI team under my guardrails.
- TypeScript
- Node.js
- InfluxDB
from 75% to 0%Dashboard maps rebuilt after an 11-minute absence
Next.js · Front-End
GPU-aware orchestration of live 3D maps in an operations dashboard
A frame manager that keeps one live WebGL map per row, verifies every frame by pixel readback and falls back to local terrain offline.
- React
- Next.js
- TypeScript
from 1.7 s to 0 sMain-thread rasterization per scenario
Re-architected how a validated spill model reaches the screen: physics in a Web Worker, geometry built once, playback paced by the front.
- JavaScript
- CesiumJS
- WebGL
from 41 to 0Proposals closed before review (of 68)
Scores signal trust, checks each segment against hydraulics and sends every new fault signature to human review before operators see it.
- TypeScript
- Node.js
- Next.js
from 0% to 99%Frames on time at 10× playback
Next.js · Front-End
Full-stack performance engineering for a real-time monitoring platform
Audited the web client, API and 3D map service end to end: 31 measured findings, 25 shipped, 6 rejected on evidence, 10× replay on time.
- React
- Next.js
- MUI
from 4.3× to 1×Drawn volume ÷ released volume
An auditor that follows the mass from the leak model to every drawn frame: three error layers fixed, four invariants enforcing balance.
- TypeScript
- Node.js
- CesiumJS
from 14 to 0Client demo views failing a walkthrough (of 38)
Moved a 3D twin to its own cloud host with managed settings and caching, and brought its client demo level with the product, checked daily.
- Azure App Service
- Azure CLI
- Node.js
from 20,000 ms to 74 msWorst-case wait from click to row update
React · Front-End
A unified alarm-acknowledgement workflow for control-room operators
One acknowledgement path for four entry points, with server-confirmed row states and one shared count: feedback in 74 ms, not up to 20 s.
- TypeScript
- React
- Next.js
from 34% to 100%API routes with a server-timing breakdown
React · Front-End
Performance instrumentation and budgets built into a web client
A development-only monitor that traces every page, API route and chart, judges each KPI once and names the file at fault; 0 bytes ship.
- React
- Next.js
- TypeScript
from 3 to 9Solved scenarios playable offline
Opened an analyst-only spill model to engineers: nine solved scenarios on real terrain, fully offline, every input bounded by physics.
- JavaScript
- CesiumJS
- Node.js
from 3 to 11History requests answered on a 24 h or 7 d view (of 11)
An admin-only console on the production site: 22 thresholded health KPIs, server-side session tracking and private usage analytics.
- TypeScript
- React
- Next.js
from 25 to 0Values baked into the build
Each offline install reads every setting from one file at startup: no rebuilds, no values baked into code, and a build guard on what ships.
- Next.js
- TypeScript
- Node.js
from 100% to 25%Rows to scan on the longest system (% of the old table)
Generates each line's live diagram from its configuration: state as a word and a color first, validated values, zero added API load.
- TypeScript
- React
- Next.js
from 0.6 to 0.84OCR confidence score (0 to 1)
Computer Vision · ML
Checking equipment investments against 50,000+ legacy drawings
Turned 50,000+ legacy CNC drawings into evidence for a capital decision: most cutting tables were oversized for the work they do.
- Python 3.8+
- NumPy
- Pandas
from 100% to 20%Data processing time (100% = before)
Deep Learning · ML
Machining analytics: 392 cutting programs profiled for improvement
One analytical view of 20 months of machining logs: 392 program profiles and the material waste hidden in thickness returns.
- Python
- Pandas
- NumPy
from 100% to 60%Time to identify and address delays (100% = before)
JS · Front-End
A heat map that shows where manufacturing is delayed
An interactive heat map of manufacturing delays by region and order status, so operations can see bottlenecks and act 40% sooner.
- HTML5
- CSS3
- JavaScript
Serverless · Back-End
A live multi-machine timeline for laser cutting jobs
A shared, live planning board for laser cutting that replaced manual scheduling across several machines, serverless end to end.
- AWS CDK
- TypeScript
- AWS Lambda
Automated intake of PDFs and machine logs into a prediction model, with OCR, ETL and monthly retraining, replacing manual work.
- S3
- DynamoDB
- Textract
Rewrites a resume into one focused page through a guided chat, with a validation layer that checks the output against the original.
- React
- Amazon S3
- AWS Amplify
Serverless · Back-End
Live production cost per machine and product, from machine logs
Live cost per machine and per product from laser, CNC and welding-robot logs, with forecasts of completion time and total job cost.
- AWS Step Functions
- React.js
- Streamlit
React · Front-End
Live production view with completion-time forecasts
A live view of three critical machines with per-machine drill-downs, completion-time forecasts and alerts that protect delivery dates.
- React.js
- Streamlit
- Python
02Stack
The stack, by use.
Most used · by project count
- Node.js13 projects
- Next.js11 projects
- React10 projects
- TypeScript10 projects
- CesiumJS9 projects
- Python7 projects
- DynamoDB5 projects
- InfluxDB5 projects
- JavaScript5 projects
- Pandas5 projects
Certifications
- TensorFlow Developer Professional Certificate
- Deep Learning Specialization
- AWS Solutions Architect
- Machine Learning Engineering for Production (MLOps) Specialization
- Practical Data Science on the AWS Cloud Specialization
- Reinforcement Learning Course - University of Alberta
03About
One architect, from data to 3D view.
Experienced data scientist with a proven track record of optimizing production processes in manufacturing using machine learning and cloud technologies. Delivers actionable insights, improves efficiency, and reduces downtime.
Draw the whole system first
The right fix shows up in the diagram.
Keep 3D instant
Off the main thread, and idle when idle.
Prove it on the running system
Evidence, not estimates.
end-to-end · data → 3D view
- 01Data
Sensors, logs, documents
- 02Models
Machine learning and physics
- 03Services
APIs and rule engines
- 04Cloud
Deploys and observability
- 053D view
Live maps in the browser
04Contact
Planning a 3D or spatial system?
Message me on LinkedIn.