# AGI Foundations for AWS (2025/05 - present)

# AGI Foundations | Speech Inference (2024/09 - 2025/05)

# AGI | Automatic Speech Recognition (2024/05 - 2024/09)

# Alexa | Connected Devices (2020/08 - 2024/04)

Non-profit dedicated to expanding access to educational microbiology projects through an open community lab & online open-source material

● Spearheaded and developed the entire Python / Flask API and Backend for the BioArt Bot project, utilizing an Opentrons OT2 robot to create real-life images out of modified strains of colorful e-coli bacteria

● Launch of BioArt Web App and API has increased scope of user-base from restrictive Bay Area bio-hackers, to enthusiasts around the world (with commissions from 10+ different countries already)

● Extended the Python / Flask API to manage the internal workflow of printing/growing the images, cutting onboarding time by 3 hours, and also automated user notifications through email, removing previous manual processes

● Managed milestones on GitHub to facilitate contributions, and simplified the dev-cycle by containerizing the application in Docker, which 2x the number of open-source contributions

Development and research of virtual machine technologies as part of the University of Manchester’s Advanced Processor Technologies Research Group

● Parallelised optimizing compiler for Maxine VM (a state of the art JVM, written in Java, and built for research/experimentation), resulting in 3x speedup across all benchmarks

● Fixed major bug in Tornado VM (a framework for executing Java programs on heterogeneous devices such as GPUs and FPGAs) that hung the JVM when executing in parallel and had prevented the team from providing async execution of framework for 6 months+

● Built an asynchronous thread-safe profiler/logger with pluggable output format (built-in json). Incurred 7% overhead (requirement <10%) for Tornado VM benchmark that produced >25MB/second of profiling data

Manufacturer of scientific detector systems covering the range of visible to x-ray

● Redesigned housing for infrared camera, optimizing for manufacture and assembly, which reduced unit-cost by 16% on an item that generates £100,000+/year in revenue

● Assembled cameras from scratch, printing circuit boards, soldering connectors, and assembling components