Scarlett Scott

Full stack software engineer with five years of experience in developing and implementing technical solutions. Analytical problem solver with communicative ability to explain complex topics to non-technical colleagues

• Web application where users can purchase other users posts, scheduled live classes or subscriptions.
• Django Rest Framework and Channels, React, Redis, Stripe, AWS S3, Sockets, Peer to Peer, Postgres. All media content is managed by AWS S3 buckets.
• All payments are handled through Stripe utilizing their payment intents and subscriptions system.
• Instant messaging via sockets, Django channels and Redis. Video calling and live classes using Reacts Peer to Peer Connections packages.
• Complex search, weighted and ranked using Django's Trigrams and vectors. Allows users to search through posts and other users profiles.

• Full stack application including asynchronous local home server. Camera security system users can view remotely featuring phone alerts when an unknown person is on the property. Features object detection and facial recognition.
• Websockets, Django Rest Framework and Channels, React, Redis, Twillio, AWS S3 Raspberry Pi, ESP 32 Cam. Libraries: OpenCv, requests, asyncio, multiprocessing
• Camera units connect to custom asynchronous local websocket server where their data and socket is stored in a map using their unit ID as the key for faster retrieval of unit information.
• Camera frames are processed with OpenCv object detection. If a person is present and their facial recognition encodings cannot be found in allowed persons, an alert is sent via text message to head of household. The frame containing the intruder is saved and uploaded to an AWS S3 bucket, a record of the incident is saved in the applications database.
• Remote Viewing: user sends ping from frontend through remote server to the local server via sockets containing the ID of the unit they are wanting to view the video stream of. Frames are then forwarded individually from the local server over sockets to the user. This ensures that the cameras ip address is never public and reduces the risk of someone unwanted viewing the streams.