Paul Tsochantaris is a Senior Software Engineer with 12+ years of experience building mobile, desktop and embedded systems, currently shaping product engineering at Car & Classic and running independent projects through bru.build. He has a strong track record bootstrapping and scaling iOS apps at companies like Deliveroo and Peanut, and has delivered everything from simple consumer apps to mission-critical industrial systems with multi-year lifecycles. Paul combines hands-on app development with low-level performance engineering—his open-source contributions include Metal backend optimizations and memory fixes for high-profile C/C++ projects such as llama.cpp and whisper.cpp. He is comfortable across the stack, from UI refinements and feature work on iOS to debugging and refactoring GPU compute code for improved efficiency. Based in Waverley, UK, he brings pragmatic engineering judgment informed by academic training in computer science and telecommunications. A detail-oriented problem solver, he often finds and fixes subtle memory and resource-management bugs that others miss.
12 years of coding experience
12 years of employment as a software developer
Hellenic College of London
Master of Science (MSc), Computer Systems Networking and Telecommunications, Master of Science (MSc), Computer Systems Networking and Telecommunications at Westminster University
ACS Athens
Bachelor of Science (BSc), Computer Science, Bachelor of Science (BSc), Computer Science at Queen Mary, U. of London
Managing Pull Requests and Issues For GitHub & GitHub Enterprise
Role in this project:
iOS Mobile Developer
Contributions:76 releases, 1 review, 668 commits in 8 years 11 months
Contributions summary:Paul primarily contributed to the development of the iOS mobile application "Trailer". Their work focused on fixing issues related to switching iOS tabs and simplifying code for image loading. The user implemented features for filtering specific properties and implemented support for the display of base and head branches for pull requests. They also made UI adjustments.
Contributions:2 reviews, 11 PRs, 5 pushes in 10 months
Contributions summary:Paul primarily contributed to the `ggml-metal` backend of the `llama.cpp` project, focusing on optimizing Metal performance and addressing memory-related issues. Their work involved debugging, refactoring, and improving the efficiency of compute operations within the Metal framework. Key contributions include fixing memory leaks, optimizing command buffer management, and removing unnecessary code. They also worked on localizing logic within the `ggml_metal_graph_compute` function for improved performance.
cppllm-inferenceggml
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.