•Headed ESP32 firmware development for Smart Nora's new Pebble 2.0, a machine learning IoT anti-snoring device
•Engineered the product's entire data flow architecture - from audio generation to serial bus transmission, to post-processing in the cloud
•Built an AWS pipeline that communicated S3 bucket events through SQS queues and CloudWatch alarms to direct AutoScale cluster scaling policies - reducing costs by 90%
•Utilized C sockets to build an IETF compliant DNS server on the ESP32 - allowing users to connect via a captive portal
•Pioneered a protocol for inter-processor communication over an SPI serial bus to send & receive audio data at a throughput of 192 kBit/s
•Eliminated dropped audio data packets through a custom SHA-256 encoded multi-part upload scheme over TLS
•Constructed a debugging and deployment pipeline utilizing SEGGER J-link tools - halving test time
•Halved snore detection classifier debugging times through the creation of CLI plotting tool using matplotlib
