Senior Software Engineer specialising in cross-language SDKs and high-performance Python and native systems. Builds developer platforms and execution systems that improve reliability, reproducibility, and cross-language consistency across large-scale codebases.
Developed SDKs and tooling for a real-time data platform, improving correctness, reproducibility, and cross-language consistency across client libraries.
Built an AST-based transpilation pipeline (C# → Python) that automated the majority of manual SDK porting effort, ensuring consistent, idiomatic, type-aware implementations across languages.
Introduced a DSL-driven intermediate representation that standardised behaviour across multi-language SDKs, reducing divergence and improving long-term maintainability.
Architected a unified cross-platform build and developer tooling platform adopted across core repositories, removing environment-specific failures and improving reproducibility and build reliability across the engineering organisation.
Integrated LLM-assisted test-suite translation into deterministic transformation pipelines, enabling automated porting while preserving correctness and reproducibility guarantees.
Exposed native CBOR and delta-calculation logic to Python via PyBind11 and PyO3, ensuring parity with underlying C/C++ behaviour without reimplementation.
Introduced OpenTelemetry instrumentation across the SDK, enabling end-to-end tracing and improving observability of distributed behaviour.
Diagnosed and resolved performance-critical issues using LLDB, GDB, Valgrind, and sanitizers.
Migrated an ad-hoc FogBugz-based Git hook system to a rigorously tested Python implementation (using property-based testing to cover Git graph state space), integrated with JIRA and enforced across all engineering repositories, improving consistency and developer experience.
Parallelised Python SDK test/build orchestration with a resource-aware scheduler driven by CPU and memory utilisation, significantly reducing build times while preventing resource contention.
Contributed to the Python SDK and its integration with the libcouchbase C/C++ client, improving observability, correctness, and cross-language runtime behaviour.
Integrated distributed tracing into the Python SDK via native bindings, improving visibility into client behaviour in distributed systems.
Developed Python ↔ native interop layers (ctypes/CPython), resolving lifecycle and memory ownership issues and improving runtime correctness.
Diagnosed and resolved performance-critical and memory issues using LLDB, GDB, Valgrind, and sanitizers.
Delivered fixes and enhancements to libcouchbase, improving behaviour exposed through higher-level SDK APIs.
Maintained build, packaging, and test infrastructure across environments, ensuring reliable releases and consistent behaviour.
Worked on NLP pipelines and C++-based transformation systems for text normalisation in TTS products.
Developed a high-performance C++ text normalisation DSL and runtime used by linguists across embedded devices and large-scale production pipelines.
Converted a legacy C++ component into a fully tested set of Java services, improving reliability, flexibility, and operational scalability while processing millions of messages daily.
Built analysis and evaluation tooling to improve linguistic coverage and system quality.
Extended schema-driven processing pipelines, improving validation and transformation reliability.
Refactored legacy processing systems into scalable service-based architectures.