6 more images
|
PickupMVP
Play & reward yourself, your community, & all athletes in any sport.
A web and mobile application to find, create, promote and win rewards from live sports matches. Available in the
App Store &
Google Play Built front-end, backend, and AI for PickupMVP.com
- Wrote abstractions for Track Authority Music (TAM) to support Groups within Groups while inheriting
publish and privacy settings from parent group
- Added customized Checkout process to require Release of Liability agreement
- Extended e-commerce system to allow draft orders and entrance fee payments of upcoming matches
- Enhanced TAM permissions matrix to allow connecting music playlists from other groups as background
music for upcoming match.
- Built Flutter WebView wrapper to publish as mobile app in App Store and Google Play
- Trained a neural net model for detecting Balls and Players in video recordings
- Wrote all business logic for Object Tracking and isolating the clip's subject athelete and ball
- Built predictions review and feedback interface to allow additional tagging and fine tuning
- Configured all server architecture to store clips on a mounted disk outside of the runtime VM
|
22 more images
|
Cosmo
Bringing order to Cyber Operations
A cyber threat intelligence platform built with the cyber planner in mind.
- Rebuilt OpenCTI interface to follow streamlined workflow to control ingestion, triage alerts and fine tune ML engine analysis
-
Wrote all parsers for transforming emails and various log formats into appropriate STIX2 objects
-
Built ingestion system to support custom configurations and toggle persistent storage of Observables while preserving analysis and prediction with or without historical data.
-
Wrote test cases and reproducible pen test with ZAP Security Scanner to validate performance, processing and predictions.
-
Used Stix2 ANTLR4 grammar for pattern matching log data with existing Indicators of Compromise
-
Leveraged inference engine settings to automated indirect relationships between different Stix objects
|
|
Imagerie
A TensorFlow training interface
Automate quality assurance with image detection
- Built interface and middleware components to robust automated image training program
- Express Server as proxy service to upload / download photos from Minio / AWS S3 and other storage services
- Drop and Drop photo library manager for building trainings from remote and local photos
- Customized Tree component for infinite nesting of tags under photos, and easier review of tag and photo collections with counts and average region sizes
- Customized drawing component for creating, dragging, and resizing tags with 0 - 400% magnified zoom window
- User defined categories, tag names, and tool sets
- Automated demo builder with customizing parameters
- Webpack 4 hotloading development environment with testing, production and development build scripts for packing and deployment
|