Arshbeer Singh Email & Phone Number

A proposal for enhancing network communications by re-implementing legacy support for IPv4 and IPv6 in v4.15.0 was approved and executed.Significant contributions made to the v4.16.0 BACnet firmware (Red5 product line) include implementation of:• BACnet/Wi-Fi and Mesh (IEEE 802.11s)• PG Python (Custom Python Interpreter and Compiler for BACnet)• Network Whitelist for BACnet/Ethernet, BACnet/IPv4, and BACnet/Wi-FiIn research and development, instrumental contributions include researching and prototyping:• Instant Firmware Transfer for v4.16.0 firmware, which reduced firmware loading time by approximately 95% (down from 7 minutes to 15 seconds) by implementing a patching-based upgrade process• Power over Wi-Fi (PoWiFi) for Red5 product line to make BACnet controllers completely portable by transmitting data and power over Wi-Fi• an Augmented Reality based Commissioning Tool for v4.17.0 firmware

British Columbia, Canada

Significant contributions made to v4.14.0 BACnet firmware (eBUS, IRC, and Red5) include:• implementation of a BACnet/MSTP Network Whitelist for filtering devices on the subnet based on MAC addresses or BACnet device instances• prototyping, proposing, and re-implementing the legacy BACnet database upgrade process, which resulted in an 80% reduction in time consumed (from 17 minutes to 3 minutes)Regular bug fixes and enhancements were also made to IPv4 and IPv6, as well as the Modbus driverIn research and development, instrumental contributions include researching, prototyping, and demoing:• BACnet Hardware and Firmware Redundancy Control for the Red5 product line• PG Python (Python support for BACnet PG objects) for v4.16.0 firmware• BACnet/Wi-Fi and Mesh (IEEE 802.11s) for v4.16.0 firmware• a GCL+ to Python translator using ANTLRv4 to eliminate the overhead of maintaining GCL support internallyAdditionally, research was conducted and a prototype was created for integrating a speech and hand gesture recognition module within an eZNT (enteliZone Network Thermostat) using the existing I2C communication bus.

British Columbia, Canada

Waterloo, Ontario, Canada

Proposed and collaborated with a team of SFU researchers and Google Engineers on an ambitious 8-month academic project that set out to exploit the advantages of integrating machine learning models into communication protocols.The project involved extensive research, prototyping, implementation, experimentation, and finally, the release of an innovative self-learning communication protocol - ConneQT.Our team successfully developed this universal communication layer that can be easily deployed on any Linux-based device, acting as a comprehensive, all-in-one solution for all your communication needs. Stay tuned for more detailed insights, which will be made public soon.

Helsinki, Uusimaa, Finland

Provided short-term software maintenance and integration services for a machine learning-based bookkeeping and error management tool that I had the privilege of selling to the organization.As part of my responsibilities, I seamlessly integrated the tool into their existing products, conducted collaborative discussion meetings with software developers and maintainers, and provided them with comprehensive implementation design plans to ensure a smooth transition and enable them to independently enhance the tool in the future.

Seattle, Washington, United States

Privileged to have worked with the oneAPI core development team on a 4-month long co-op opportunity.